Spatial variation of dissolved organic matter in the Tuojiang River Basin in Chengdu, China: insights based on EEMs–PARAFAC analysis

A sound ecological environment is one of the most important contributors to a population's wellbeing. The Tuojiang River Basin – a major first-level tributary that flows into the upper reaches of the Yangtze River – is an essential support point for the Yangtze River protection strategy in China. Its water environmental conditions have a direct influence on the incoming water quality in the middle and lower reaches of the Yangtze River. In recent decades, the dual impacts of sociodemographic and economic development have exerted tremendous pressure on the water ecological environment of the Tuojiang River, making it the most seriously polluted river basin in Sichuan Province.

Despite extensive efforts toward water quality improvement in recent years, parts of the Tuojiang River are still heavily polluted (Fan et al. 2022). Persistent problems in this area are exemplified by seasonal overloading of water environmental capacity and failure of water ecological function, which raise serious environmental issues. In particular, discharge of urban domestic sewage as well as industrial and agricultural wastewater has a strong impact on water quality in the Tuojiang River Basin in Chengdu, leading to considerable water quality differences between the main stream and tributaries (Qin et al. 2020). The overall water quality of the main stream is satisfactory, in contrast to the much poorer water quality of its tributaries. Black and odorous water was generally eliminated, despite potential risk of recurrence in some small tributaries. Given the direct influence of water quality on the life and productivity of urban residents, addressing the problems of water environmental pollution in the Tuojiang River Basin in Chengdu is imperative.

Dissolved organic matter (DOM) is the largest organic carbon pool in aquatic ecological environments. It plays a significant role in maintaining the carbon cycle in aquatic ecosystems, as well as the degradation and adsorption of heavy metals and organic pollutants (Wu et al. 2010; Liu et al. 2019). DOM is widely present in surface water and groundwater as a consequence of different hydrological, biological, and geological interactions (Leenheer & Croue 2003; Mayayorga et al. 2005). It supports heterotrophic microbial metabolism (Stedmon et al. 2011) and, to some extent, affects the integrity and function of riverine ecosystems. The composition of DOM is determined by its source. Riverine DOM can be derived from autochthonous inputs (e.g., phytoplankton, aquatic plants, and heterotrophs), allochthonous inputs (e.g., soil leaching, rock weathering, and atmospheric deposition), and anthropogenic inputs (e.g., industry, wastewater, intensive agriculture, and farms) (Fisher et al. 2004; Elliott et al. 2006; Giorgio & Pace 2008; Griffith & Raymond 2010; Fashing et al. 2014).

Three-dimensional excitation–emission matrix fluorescence spectroscopy coupled with parallel factor analysis (EEMs–PARAFAC) is widely used for DOM source analysis. In this approach, the specific identification of organic compounds is performed using excitation and emission matrices (Zheng et al. 2016). It has advantages in terms of high sensitivity and fast detection time in tracking the quality and quantity of natural DOM (Stedmon et al. 2006; Mohammad et al. 2009). PARAFAC resolves the fluorescence spectra of DOM to obtain the source and characteristics of different components (Zhang 2019). Using EEMs–PARAFAC, Zhang et al. (2022) traced the source and composition of DOM in black and odorous rural river water, where the DOM contained six fluorescent components under the influence of both autochthonous and allochthonous sources. Additionally, Chen et al. (2022) explored the influence of different flooding conditions on the source of DOM in lake sediments based on EEMs–PARAFAC. They reported that protein-like component accounted for a higher proportion of DOM under seasonal flooding conditions, whereas humus-like components exhibited the opposite pattern. Furthermore, Xu et al. (2022) analyzed the spatial variation of DOM in river water by EEMs–PARAFAC and observed remarkable spatial variation of water quality parameters. As an important carrier of pollutants, DOM is closely related to the migration and transformation of pollutants, bioavailability, and nutrient retention and release (Smith et al. 2021; Wen et al. 2021; Li et al. 2022). Therefore, the source, characteristics and other information of DOM can be judged according to different components identified by the water body DOM. Accordingly, these previous studies have demonstrated the effectiveness of EEMs–PARAFAC in the analysis and identification of DOM sources and its compositional characteristics.

Herein, EEMs–PARAFAC was adopted to characterize the variation of DOM characteristics along the Tuojiang River Basin in Chengdu as well as its main stream and tributaries. The aim of this study was to determine the sources of DOM and its material composition in the river water. The results provide useful information to guide source control, sewage interception, and water environmental quality improvement in the Tuojiang River Basin.

The Tuojiang River is the most important river in Sichuan Province and a major tributary on the left bank of the Yangtze River. The Tuojiang River Basin (103°41′–105°55′ E, 28°50′–31°41′ N) is adjacent to Jiufeng Mountain in the Longmen Mountain range in the northwest and neighbors the Minjiang River in the west. It connects to the Fujinag River in the east and flows into the Yangtze River in the south. The Tuojiang River originates in Dahei Bay under fault rockhead in Mianzhu City, at the southern foot of Jiuding Mountain. It flows through the cities of Deyang, Chengdu, Ziyang, and Neijiang, and then discharges into the Yangtze River through Guanyizui in the urban area of Luzhou City. The basin area (in Sichuan) is 2.56 × 10⁴ km², accounting for 5.3% of the total provincial area. The study area is part of the Tuojiang River Basin, comprising the administrative areas of Chengdu, Deyang, Ziyang, Mianzhu, Guanghan, and Jinyang (hereinafter referred to as the Tuojiang River Basin in Chengdu). The Tuojiang River Basin is the area with the most concentrated towns, the most densely populated and the strongest economic strength in Sichuan Province. The economic development in the basin is rapid, with densely covered towns and complete industries of agriculture, forestry, animal husbandry, and fishery. The red layer hilly areas in the middle and lower reaches of the basin (such as Jianyang, Zizhong, Anyue, and other rural areas) are dominated by traditional fragmented farming methods, with high use of pesticides and fertilizers and low utilization rate and serious pollution from agricultural non-point sources. Industry in the basin is developed, such as Deyang heavy machinery manufacturing.

The main experimental instruments were a HACH portable multi-parameter water quality analyzer (HQ30d, Shenzhen, China), which was used for in situ measurements of water temperature, dissolved oxygen (DO), pH, and oxidation–reduction potential [ORP]; a nitrogen gas cylinder; a biochemical incubator; a DR6000 UV–Vis spectrophotometer (Shenzhen, China); a DRB200 digestion system (Shenzhen, China); a total organic carbon analyzer; an ultrapure water system; and pipettes.

After sampling, permanganate index (COD_Mn), ammonia-nitrogen (NH₃–N), total nitrogen (TN), and total phosphorus (TP) were determined according to Analytical Methods for Water and Wastewater Monitoring (State Environmental Protection Administration 2002a). Briefly, COD_Mn was analyzed using the potassium dichromate method; NH₃–N was quantified using Nessler's reagent spectrophotometry; and TN and TP were determined by alkaline potassium persulfate digestion–UV spectrophotometry and ammonium molybdate spectrophotometry, respectively. Dissolved organic carbon (DOC) was determined using a total organic carbon analyzer (TOC-L CPN, Shimadzu, Japan). Three parallel samples were analyzed for each measurement.

A fluorescence spectrophotometer (RF-6000, Shimadzu, Japan) was used to analyze three-dimensional fluorescence spectral characteristics of the DOM. To eliminate the internal filtering effect, samples containing >8 mg DOC/L were diluted with Milli-Q water to a concentration of 8 mg DOC/L before analysis. Mill-Q ultrapure water was used as a blank. The excitation wavelength (E_x) ranged from 200 to 400 nm in 5-nm intervals and the emission wavelength (E_m) ranged from 250 to 450 nm in 2-nm intervals. The scanning speed was 2,000 nm/min.

The sampling point distribution map was drawn using ArcGIS 10.2 (Environment System Research Institute Inc., Redlands, CA, USA). Data processing and correlation analysis were performed using Origin 2022 (study version; OriginLab Corp., Northampton, MA, USA) and Excel 2016 (Microsoft Corp., Redmond, WA, USA). PARAFAC analysis of EEMs was carried out using the DOMFluor toolbox in MATLAB 2018b (MATLAB, MathWorks Inc., Natick, MA, USA), and the results were subjected to cluster analysis using SPSS 26.0 (IBM Corp., Armonk, NY, USA).

The values of major water environmental parameters in the Tuojiang River Basin in Chengdu are provided in Table 2. The mean parameter values of DO (8.23 mg/L) and pH (7.61) across the sampling points met the Class I water quality standard according to the Environmental Quality Standard for Surface Water (State Environmental Protection Administration 2002b). The range of water quality category in GB 3838-2002 is: Class I: DO ≥ 7.5 mg/L; Class II: DO ≥ 6 mg/L; Class III: DO ≥ 5 mg/L; Class VI: DO ≥ 3 mg/L; Class V: DO ≥ 2 mg/L. The DO values at some sampling points, such as 10#, 14#, and 18#, reached Class III water quality standards, but the water quality at tributary sampling points 4# and 5# was still poor after converging into 7#. Additionally, poor water quality of Class V and lower was recorded in the main stream sampling points 11#, 12#, and 20# as well as at the tributary sampling point 17#. The DO values at all other sampling points met Class I water quality standards, and the highest value of 12.12 mg/L was recorded at point 13#.

The DO and ORP values show the coefficients of variation between 10 and 90%, indicating its moderate spatial variability. The maximum ORP of 444.9 μS/cm was observed at point 4#. The pH and temperature values show relatively small coefficients of variation, indicating minor spatial variation in these two parameters across the sampling points, with no significant differences between the main stream and tributaries.

Both COD_Mn and DOC values are greater in the main stream than in the tributaries, which mean that the degree of organic pollution is more serious in the main stream than in the tributaries, and there are significant differences between the main stream and tributaries. DOC varies from 0.55 to 9.14 mg/L across the sampling points. It has been reported that higher DOC values occur in water bodies in areas strongly influenced by anthropogenic discharges (Huang et al. 2016a, 2016b). In the present study, point 17# is close to an agricultural wastewater discharge channel, leading to it presenting the highest DOC value (9.14 mg/L). In contrast, point 1# is located in a mountainous area with low water flow and sediment deposits, leading to it presenting the lowest DOC value (0.55 mg/L). Owing to the strong influence of human activities at 201 Hospital, point 7# presents a higher DOC value than its adjacent sampling points. After tributary convergence, the DOC values exhibit a decreasing trend along the main stream flow direction. Refer to Figure 3 and Table 2 for specific values.

The SUVA₂₅₄ and S_R values at various sampling points show significant differences between the main stream and tributaries. The parameter values are higher for the tributaries than for the main stream, indicating a higher degree of DOM humification and aromatization as well as greater autochthonous contribution in the tributaries than that in the main stream. The SUVA₂₅₄ values vary from 0.30 to 3.51. Compared with other sampling points, the SUVA₂₅₄ values at points 1#, 2#, 3#, and 18# are remarkably larger, and the higher degrees of aromatization and humification are due to the strong influence of human activities. DOM aromatization in the tributaries decreases along the flow direction. S_R values range from 1.25 to 6.35 (all > 1), indicating that the DOM in this water body is mainly from autochthonous sources. At point 7#, the autochthonous input is relatively low compared with that at other sampling points, which may be due to its proximity to a sewage outlet and low DO concentration.

The biological index (BIX) measures the autochthonous characteristics of DOM in water bodies, which are proportional to each other (Chen et al. 2017). The results indicate both autochthonous and allochthonous sources for DOM in the Tuojiang River Basin in Chengdu, its main stream, and tributaries. The BIX values range from 0.72 to 1.01 (Figure 7), indicating that the DOM sources at different sampling points in the water bodies are influenced by both terrestrial and autochthonous inputs. The autochthonous characteristics of the DOM are stronger in the main stream than in the tributaries. The BIX value at point 6# (0.72) is remarkably lower than those at other sampling points. The DOM at point 6# mainly originates from terrestrial inputs and is strongly influenced by nearby industrial wastewater discharge. The BIX value is highest at point 8# (1.01), where autochthonous input is most prominent, which is consistent with the FI results and those shown in Figure 6(a).

Generally, the humification index (HIX) is used to reflect the degree of DOM humification (Bu et al. 2019). A higher HIX value indicates a higher degree of DOM humification and a stronger terrestrial origin; a lower HIX value suggests a stronger autochthonous origin of the DOM. The HIX values are basically < 4.0 in the Tuojiang River Basin in Chengdu, its main stream, and tributaries (Figure 6(b)), which indicates a relatively strong autochthonous origin of the DOM in the water bodies. In contrast with other fluorescence parameters, HIX fluctuates more prominently in the range 1.40–5.02 (Figure 7). The degrees of DOM humification at points 17# and 18# are higher than those at other sampling points owing to the strong influence of industrial wastewater and domestic sewage discharge, respectively. This indicates a strong terrestrial origin of the DOM at the two sampling points (17# and 18#). A previous study has shown that farmland, woodland, and grassland near sampling points can lead to the enhancement of terrestrial origin, aromaticity, and humification degree of DOM in water bodies (Liu 2018). There is a relatively high degree of DOM humification at point 18#, supporting the SUVA₂₅₄ and FI results. The mean HIX values at the other sampling points are <4, suggesting the dominance of autochthonous DOM. Additionally, the HIX values are higher for the tributaries than the main stream. Accordingly, there is a higher humification degree of DOM in the tributaries than in the main stream, which mirrors the FI and BIX results. The autochthonous origin is weaker in the tributaries than in the main stream, similar to the results shown in Figure 6(b).

The freshness index (β:α) reflects the proportion of newly produced DOM in the overall DOM: the relative contribution of endogenous substances to DOM (Huang et al. 2016a, 2016b), with its level and variation basically consistent with those of BIX. In this study, the β:α values range from 0.69 to 0.95 (Figure 7), which indicates high biological activity and the high autochthonous characteristics of the DOM, which is in basic agreement with the results of Zhang et al. (2022). The β:α value at point 6# is relatively low, suggesting a small proportion of newly produced DOM with the weakest biological activity and autochthonous characteristics, corroborating the BIX results. This sampling point (6#) is close to a commercial area and therefore strongly influenced by terrestrial sources. The highest β:α value is observed for point 8#, reflecting the strongest biological activity and most prominent autochthonous characteristics of the DOM at that point, which mirrors the FI and BIX results.

The composition of the DOM fluorescence in the Tuojiang River Basin in Chengdu as well as its main stream and tributaries were analyzed by PARAFAC analysis. In total, four components of two types were identified in the basin and its tributaries. Comparison of component type with those reported by Coble et al. (1990) revealed that they are humic-like (C1, C2, and C4) and protein-like (C3) components. Additionally, three components of two types were identified in the main stream, which are humic-like (C1 and C2) and protein-like (C3) components (Figure 8 and Tables 2–7 of of supplemental materials).

The fluorescent component C1 (255/426 nm) in the basin and its tributaries corresponds to fluorescence peak A, similar to the pattern of a typical UV fulvic acid-like peak. This component is associated with relatively high aromaticity and high molecular weight DOM groups, which are resistant to biodegradation and indicative of terrestrial input (Yan et al. 2021). C2 (260, 345/454 nm) in the basin and C2 (260, 345/462 nm) in the tributaries match fluorescence peak F, which resembles the pattern of a typical humic acid-like peak. Such components originate primarily from domestic sewage and represent a typical terrestrial humic-like substance (Garcia et al. 2018). C3 (235, 305/338 nm) in the basin and C3 (235, 305/342 nm) in the tributaries correspond to fluorescence peak T, which mirrors the pattern of a typical tryptophan-like peak. Such components are prone to biodegradation, strongly influenced by human activities (Yi et al. 2017), and derived mainly from urban domestic sewage and food industry wastewater (Francisco et al. 2020). C4 (265/482 nm) in the basin and its tributaries matches fluorescence peak F, similar to the pattern of a typical humic acid-like peak.

In the main stream, the fluorescent component C1 (255/418 nm) corresponds to fluorescence peak A, similar to the pattern of typical UV fulvic acid-like peak. C2 (260, 340/442 nm) corresponds to fluorescence peaks F and C, which mirror the patterns of traditional humic acid-like and visible fulvic acid-like peaks, respectively. The position of the visible fulvic acid-like peak C is unstable, which may be red-shifted or blue-shifted due to differences in sample properties (Zhong et al. 2008). The DOM components may be degraded or transformed during water migration in large rivers, with bioavailability playing an important regulating role (Gan 2013). Therefore, there is no component C4 in the main stream. C3 (235, 305/334 nm) matches the fluorescence peak T, which resembles the pattern of a typical tryptophan-like peak.

• Based on PARAFAC analysis, four fluorescent components of the DOM in the Tuojiang River Basin in Chengdu and its tributaries were identified, i.e., humus-like components C1, C2, and C4 and protein-like component C3. Additionally, three fluorescent components were identified in the main stream, i.e., humus-like components C1 and C2 and protein-like component C3. The spatial variation of DOM concentration in water bodies is large. Correlation analysis revealed a homology for the four DOM fluorescent components in the basin and its tributaries. In the main stream, DOM component C1 is derived from the same sources as C2 and C3, although the latter two components show different sources.

• The UV absorbance parameters (SUVA₂₅₄, S_R) and fluorescence parameters (FI, BIX, HIX, β:α) indicate that both autochthonous (mainly from soil, human activities, surface runoff and decomposition of plant and animal residues) and allochthonous (formed by biological activity in water bodies) inputs contribute to the sources of DOM in the Tuojiang River Basin in Chengdu, with a greater autochthonous contribution. This can be used to judge the source of DOM and provide a basis for the treatment of Tuojiang River Basin. Based on FI, HIX, and component proportions, the DOM shows stronger autochthonous characteristics in the main stream than in the tributaries, which is primarily due to the strong influence of urbanization on tributary waters.

• In the Tuojiang River Basin in Chengdu, the population is dense, the degree of urbanization is large, the industry is developed, and the agricultural non-point source pollution is serious. The analysis results show that the BIX, FI, HIX, SUVA₂₅₄ values of the water DOM in the sampling points (6#, 17#, 18#) affected by the discharge of domestic sewage and industrial and agricultural wastewater are relatively large and have relatively prominent terrestrial origin. Therefore, we recommend that pollution source control standards be quantified by region and type, and that point source control standards for enterprises in key industries, urban non-point source pollution control standards, and agricultural non-point source pollution control standards are developed. Additionally, we suggest that decentralized harmless utilization of rural wastewater and rational desilting of urban rivers be promoted.

H.T.L., B.B.C., and B.C. conceptualized the study, did formal analysis, and carried out investigation; X.X.L. prepared and wrote the original draft; B.B.C., X.H.L., Y.W., L.X.H., and C.W. wrote, reviewed, edited, and supervised the study.

All the authors have approved the submission and consented for publication.

The present work is financially supported jointly by the National Natural Science Foundation of China (Grant No. U20A20316 and No. 72091511), the Natural Science Foundation of Hebei Province (No. E2020402074) and Science Fund for Distinguished Young Scholars of Hebei Province (No. E2022402064).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.