Introduction

Nature river ecosystems serve as important water resources and natural bioreactors for the degradation of pollution and are also a significant source of drinking water. Polluted river ecosystems may hamper local economic development (Qin et al. 2010; Wang et al. 2014). Dissolved organic matter (DOM) can be found in all natural aquatic environments, which represent one of the largest pools of organic carbon on Earth (Hedges, 1992). It is the main source of energy (carbon) for heterotrophic bacteria (Carlson, 2002). Chromophoric DOM (CDOM) is defined as the component of the total DOM that absorbs light in the ultraviolet and visible range of the electromagnetic spectrum (Coble, 2007). CDOM is a complex mixture of organic compounds that originates from a wide range of sources such as the decomposition of bacteria, algae, or plants and their continuous transformation through photochemical and microbial processes that plays an important role in the global carbon cycle (Coble, 2007; Kowalczuk et al. 2013). CDOM, present as humic-like components and protein-like components, represents a large fraction of the dissolved organic material that contributes to light absorption and fluorescence by non-chlorophyllous material in coastal regions (Fellman et al. 2010). CDOM can also contribute to global warming through the mineralization of dissolved organic carbon (DOC) (Bertisson and Tranvik, 2000; Zhang et al. 2011). The mineralization of carbon-containing organic compounds always leads to CO2 and CO. CDOM can be used to discriminate and trace water masses (Coble et al. 1998) and estimate the concentrations of DOC and specific constituents (e.g., amino acids and lignin) (Guo et al. 2011).

The fluorescent characteristics of CDOM have been extensively investigated owing to the technological advances in fluorescence spectroscopy, especially the development of excitation-emission matrixes (EEMs) (Coble, 1996), and successfully applied to identifying terrestrial, marine, and anthropogenic components (Baker 2011). EEM offers several major advantages over single-scan methodologies such as providing new information regarding the fluorescence CDOM composition and source of a sample (Coble et al. 1993). However, to conquer the difficulty in identifying individual fluorescence components in water, simple and multivariate data analysis techniques such as principal component analysis (PCA) (Boehme et al. 2004) and parallel factor analysis (PARAFAC) have been used (Stedom et al. 2003). EEMs can capture miscellaneous fluorophores present in CDOM, which can also identify different characteristics of CDOM with various sources and biochemical pathways (Hudson et al. 2007). EEMs combined with PARAFAC have been applied to trace photochemical and microbial reactions with organic matter (Stedmon and Markager 2005a; Singh et al. 2010). PARAFAC analysis can decompose fluorescence EEMs into independent groups of fluorescence components (He et al. 2013) and provide a unique solution to the DOM EEM dataset, and it is regarded as an important analytical tool to characterize the CDOM in complex coastal regimes (Singh et al. 2010). Two primary fluorescent groups have been identified in CDOM in natural waters: humic- and protein-like. Humic-like substances are complex mixtures containing aromatic and aliphatic compounds derived from the decay of organic matter, whereas protein-like substances have to do with high biological activity (Harvey and Boran 1985; Stedmon et al. 2003). The number of studies on CDOM using EEM-PARAFAC in estuarine environments is increasing (He et al., 2013; Guo et al., 2011). In the environment, the variable concentration and composition of CDOM have substantial effects on the health of ecosystems through processes such as nutrient availability (Boavida and Wetzel, 1998) and light attenuation (Markager and Vincent, 2000). In the Heilongjiang River, the large quantity of humic substances confers the black color to stream water and has a significant impact on the quality of the water and carbon cycling. It is therefore of paramount relevance to improve our understanding of fate of CDOM at the high concentration of humic substances river water.

Eutrophication owing to high concentrations of phosphorus and nitrogen is a serious environment problem in estuarine systems in China. Previous studies have shown that the turnover of DOM is proposed to occur with dissolved phosphorus and nitrogen indexes (Hopkinson et al. 2002; Kragh and SØndergaard, 2004). Qualls and Haines (1991) considered that the DOM plays an important role in terrestrial and stream ecosystems owing to its being a major mode of export of phosphorus and nitrogen. Therefore, characterizing the relationships between DOM and nutrients may provide an effective management of water quality and control of eutrophication. Heilongjiang is a large complex estuarine system located in the northeast of China that directly flows into the west bank of the Pacific. There are large quantities of humics in the water of the Heilongjiang River. In the southwest of the Heilongjiang River, there are many mountains from which runoff flows into Heilongjiang carrying a large amount of DOM. Therefore, the Heilongjiang receives substantial terrestrial deliveries of DOM. The CDOM fraction of the terrestrial deliveries confers a brown color to the water, often referred to as “browning,” and it plays a central role in the carbon cycle in Heilongjiang. Furthermore, the organic pollution indicators of biochemical oxygen demand (BOD5) and chemical oxygen demand (CODMn) are at a high level in some areas of Heilongjiang. Therefore, there is a need to characterize the properties of DOM to identify the influencing indicators of water quality in the Heilongjiang River. However, to our knowledge, there have not been any detailed studies characterizing the concentration, composition, source, and properties of CDOM, despite the fact that it is an important biogeochemical factor for the carbon cycle in the Heilongjiang River. Therefore, further work is needed to characterize the CDOM in the Heilongjiang River.

In the present study, we used EEM fluorescence measurements to characterize the distribution of the CDOM concentration, composition, and sources in the Heilongjiang River. Moreover, examining the relationships between different fluorescence components and water quality parameters, we sought to understand (1) the characterization of CDOM in the Heilongjiang River, (2) the influencing factors of BOD5 and CODMn, and (3) the co-variation between CDOM and nutrients.

Methods

Sampling strategy

The Heilongjiang originates from the Hailar River and its stream length is up to 4370 km. It flows through China, Russia, and Mongolia, with its watershed area in China accounting for approximately 48 % of the whole basin area. The Greater Khingan Mountains, Lesser Khingan Mountains, and Changbai Mountains are along the Heilongjiang from northeast to southwest. Surface water samples were collected from eight tributaries of the Heilongjiang River located in northeast China. The sampling areas are as follows: EER (121° 28′ 59.8″ E, 53°19′ 55.1″ N); LGR (121° 35′ 24.2″ E, 53° 21′ 08.2″ N); HMSR (126° 39′ 59.1″ E, 51° 44′ 07.5″ N); HMR (126° 36′ 27.2″ E, 51° 39′ 54″ N); HR (127° 31′ 23.8″ E; 50° 05′ 28.7″ N); MR (131° 03′ 09″ E, 47° 41′ 12.6″ N); TJR (132° 39′ 51.9″ E, 47° 57′ 18.7″ N); WSR (134° 40′ 2.8″ E, 48° 13′ 55.8″ N) (Fig. 1).

Fig. 1
figure 1

Location of the Heilongjiang River in China and the distribution of sampling sites

Full size image

Water samples were collected using acid-cleaned Niskin bottles in June 2014 throughout these eight areas. We chose 20 stations to collect samples. After the sample collections, they were held on ice in the field (without being frozen) (Fellman et al. 2008) and immediately sent to the laboratory in darkness. Some water samples were filtered using pre-rinsed 0.45-μm nucleopore membrane filters for DOC, phosphorus, nitrogen, chlorophyll (Chl-a), and fluorescence analyses. The remaining water samples were used for the measurement of CODMn, BOD5, and particulate organic carbon (POC). The water samples were stored in a refrigerator at 4 °C, and all of the measurements were made in the laboratory within 3 days after collection.

Physico-chemical analyses

Analysis of DOC

DOC was analyzed by using a SHIMADZU TOC-Vcph analyzer. First, the machine gives the total amount of carbon (TC) in the sample by burning it and the resulting CO2 is detected by a gas analyzer. Phosphoric acid is then added into the sample to obtain the inorganic carbon (IC). Finally, the subtraction of the IC from TC produces the DOC.

Analysis of CODMn and BOD5

Titration with acidic potassium permanganate was used for the determination of CODMn, which is the equivalent amount of oxygen needed to decompose organic substances present in a sample when treated with acidic potassium permanganate. The measurement of CODMn was all based on the procedures from “Monitoring and Analytical Method of Water and Waste Water” (State Environment Protection Administration of China 2002). The samples were carried without attenuation to test BOD5 using a HACH BODTrak analyzer.

Analysis of total dissolved phosphorus, dissolved reactive phosphorus, ammonia nitrogen, nitrate nitrogen, and nitrite nitrogen

Total dissolved phosphorus (TDP) was determined by the method of phosphorus persulfate digestion. The analytical determination of dissolved reactive phosphorus (DRP) was based on the phosphomolybdic acid methodology, as modified by Murphy and Riley (1962). The method of Nessler’s reagent colorimetry was used to measure the ammonia nitrogen (NH4 +–N) water quality (MEP China Ministry of Environmental Protection of the People’s Republic of China 2009). Ultraviolet spectrophotometric method was used for determining nitrate nitrogen (NO3 −–N), and nitrite nitrogen (NO2 −–N) was measured by the method of N-(1-naphthyl) ethylenediamine spectrophotometry. These analytical methods were based on the procedures from “Monitoring and Analytical Method of Water and Waste Water” (State Environment Protection Administration of China 2002).

Analysis of POC and Chl-a

The filters for particulate organic carbon (POC) were dried for 24 h at 50 °C and stored in a vacuum dryer until analysis. All of the glassware was washed with 1 M HCl and ultrapure water and combusted at 450 °C for 6 h. All the filtration samples were washed with 1 M HCl to remove carbonate. POC was analyzed by a carbon/sulfur analyzer. To measure Chl-a, the water samples were filtered using pre-rinsed 0.45-μm nucleopore membrane filters, and the filters were dried at a low temperature. After that, the filters were thoroughly dissolved in 90 % acetone solution. The extractions were centrifuged at a speed of 3000 r/min. Finally, the absorbance of the supernatant was measured at wavelengths of 750, 663, 645, and 630 nm. The formula for Chl-a was as follows:

$$ \mathrm{C}\mathrm{h}\mathrm{l}\hbox{-} \mathrm{a}\left(\mathrm{mg}/{\mathrm{m}}^3\right)=\frac{\left[11.64\times \left({D}_{663}-{D}_{750}\right)-2.16\times \left({D}_{645}-{D}_{745}\right)+0.10\times \left({D}_{630}-{D}_{750}\right)\times {V}_1\right]}{V\times \delta } $$

where V is the volume of the water sample (L), D is the absorbance value, V 1 is the volume of the supernatant (mL), and δ is the path length of the cuvette (cm). The method of measuring Chl-a is based on the procedures for “Monitoring and Analytical Method of Water and Waste Water” (State Environment Protection Administration of China 2002).

Fluorescence spectroscopy

The EEMs were measured using a PerkinElmer LS50B fluorescence spectrophotometer with a clear quartz cuvette at room temperature. Filtered water samples were irradiated with excitation wavelengths from 200 to 490 nm at 5-nm intervals, emission wavelengths were recorded from 250 to 550 nm at 2-nm intervals, and the scan speed was adjusted to 1200 nm min−1. The methods of correcting and standardizing the EEMs were as follows (Stedmon and Bro, 2008; McKnight et al., 2001; Murphy, et al., 2013): (1) The instrument was corrected according to the manufacturer’s instructions before analyzing the samples. (2) To eliminate the inner-filter effect, the EEMs were corrected for absorbance by the multiplication of each value in the EEMs with a correction factor, based on the premise that the average path length of the absorption of excitation and emission light was one half of the cuvette length. The correction is expressed mathematically as F IFE = F Corr × 10(AEx + AEm)/2, where F IFE and F Corr are the corrected and uncorrected fluorescence intensities, respectively, and A Ex and A Em are the absorbance values at the current emission wavelengths. (3) Subsequently, Milli-Q water blank EEMs were subtracted from the sample EEMs to eliminate the Raman peaks, and then the EEMs were normalized to the daily determined water Raman-integrated area maximum fluorescence intensity (350Ex/397Em). (4) The Rayleigh and Raman scattering peaks of each scan centered on the respective scatter peaks were removed and then the excised data were replaced with interpolated data using the Delaunay triangulation method by FLToolbox developed by Wade Sheldon (University of Georgia) for MATLAB. The fluorescence intensities measured were reported in Raman units (R.U.) in this study.

PARAFAC analysis

A total of 160 EEMs of water samples from the Heilongjiang River were analyzed by PARAFAC. The PARAFAC analysis was carried out in MATLAB with the DOM Fluor toolbox, following the procedure described by Stedmon and Bro (2008). For the PARAFAC modeling, excitation wave-lengths from 200 to 220 nm were deleted from each EEM because of random data fluctuations. PARAFAC statistically decomposes the DOM three-way data into individual fluorescence components with regard to their spectral shape or number (Stedmon et al. 2003). The model was run with non-negativity constraints applied to each dimension (Murphy, et al. 2013; Stedmon and Bro, 2008). An initial exploratory analysis was performed in which outliers were identified and removed from the dataset. No samples were considered outliers. Two to seven components were computed for the EEMs. The determination of the number of components was primarily based on split-half validation and the analysis of the residuals of residual loadings (Stedmon and Bro, 2008). The concentration scores of the PARAFAC components were expressed as maximum fluorescence intensities (F max) (R.U.) for each modeled component.

Statistical analysis

Correlation analysis and t test were conducted using SPSS version 22. Origin Pro 9.1 was also used to analyze the relationships between fluorescence components.

Results

PARAFAC model results

The combined data set was split into two random halves called “Half1” and “Half2” to perform the split-half validation procedure. PARAFAC models were run for each of the Half1 and Half2 data groups. The 4-component model was successfully validated in three independent data sets to ensure the correctness of the modeled CDOM components by the PARAFAC model using the split-half validation procedure. The EEM spectra and excitation and emission loadings of the four components are shown in Table 1. Table 2 provides the excitation and emission characteristics of the CDOM components with identified and examples of matching components identified by other researchers who modeled CDOM EEMs in various marine, oceanic, and estuarine environments. Although four individual components were determined for this dataset using the PARAFAC model (Table 1), our results do not suggest that only four types of fluorophores were present in all samples or that all four components were present in the majority of the transect samples (Table 1). Other fluorescent groups are certainly present, but their influence is so weak that they cannot be distinguished from the noise. Therefore, these four fluorescence components could explain most of the variation. All components contained multiple peaks, except for component C3. Generally, these four components had single emission maxima.

Table 1 PARAFAC model output showing fluorescence signatures of the four fluorescence components. The contour plots above present the spectral shapes of the excitation and emission of the four components. The line plots present split-half validation results of the four components; excitation (left) and emission (right) spectra were estimated from the two independent halves of the dataset (red and green lines) and the complete dataset (black lines)
Full size table
Table 2 Spectral characteristics of excitation and emission maxima of the four components identified by PARAFAC modeling for the whole EEMs data set collected in the Heilongjiang watershed compared to previously identified sources
Full size table

Component C1 is composed of two peaks with excitation maxima at 240 and 320 nm with one emission peak centered at 390 nm (Table 1). These fluorescence characteristics could be categorized as the previously defined humic-like peak M (Coble, 1996). The spectral features are similar to other reported microbial humic-like components and have a lower molecular weight (Stedmon et al. 2005b; Williams et al. 2010; Yamashita et al. 2008).

Component C2 is also composed of two peaks with excitation maxima at 240 and 340 nm, and emission maxima were observed in the EEMs at 436 nm, although they were red-shifted compared to C1 (Table 1). This component was characterized as a mixture of the traditional terrestrial humic-like peaks A and C according to Coble (1996). Previous studies have demonstrated that this component is derived from terrestrial plant or soil organic compounds (Fellman et al. 2010).

Component C3 exhibits a single excitation/emission wavelength pair of <240/448 nm (Table 1) and it is typically categorized as a terrestrial humic-like component (Yamashita et al. 2008; Kowalczuk et al. 2009; Mueller et al. 2012). This component has shown to be a dominant DOM exported from the catchments during the warmer months of the year and presents its highest concentrations in forest stream and wetlands (Stedmon et al. 2003; Stedmon and Markager, 2005a).

Component C4 is composed of two peaks with excitation maxima at 280 and 400 nm and one emission peak centered at 480 nm (Table 1), which could be categorized as the previously defined peaks A and C (Coble, 1996). This component represents fluorophores that have the longest excitation wavelength and broadest excitation band, as well as the longest emission wavelength associated with a broad emission band. Such excitation and emission characteristics are associated with terrestrial organic matter that is composed of large molecular weight and aromatic organic compounds (Stedmon and Markager 2005b; Kowalczuk et al. 2013). In the Heilongjiang River, we did not detect protein-like fluorescence, and the contributions of components C1–4 were all characterized as humic-like substances.

Distribution of PARAFAC components

The spatial variations of the percentage distribution of the four PARAFAC components of each study area are shown in Fig. 2. The percent distributions of C–4 in the Heilongjiang River range from 41.8 to 53.0, 16.9 to 24.1, 11.1 to 23.1, and 14.0 to 18.5 %, respectively (Fig. 2). Component C1 generally had the highest relative distribution in the Heilongjiang River at each sample site. The high intensity of C1 indicates the impact of this type of organic matter on the Heilongjiang River. The spatial trends of the relative distribution of %C1 and %C4 were generally similar, with the highest values in HR and the lowest values in WSR. There were no significant differences in the relative contribution of C2 in most of the study areas. As for the terrestrial component C3, its spatial trend of the relative distribution was opposite from that of C1, with the highest values in WSR and the lowest values in HR (Fig. 2). Figure 3 shows the plots of components against each other for the data set and the variability in the F max of the fluorescence components. C1, C2, and C4 were almost perfectly linearly correlated (p < 0.001), but C3 was not correlated with the other three components.

Fig. 2
figure 2

Changes in the relative distributions (%) of the four PARAFAC-modeled CDOM components. Note: the sample sites from upstream to downstream are EER, LGR, HMSR, HMR, HR, MR, TGR, and WSR, respectively

Full size image
Fig. 3
figure 3

Relationships between the four fluorescence components

Full size image

Spatial distribution of water quality parameters

The mean values of the physico-chemical parameters in each study area are presented in Fig. 4. The mean values of DOC and CODMn ranged from a minimum of 16.64 ± 8.83 and 7.10 ± 1.17 mg/L in HMR to a maximum of 33.73 ± 8.76 and 27.60 ± 6.80 mg/L in HR, respectively. The mean values of the POC concentration ranged from a minimum of 5.45 ± 1.91 mg/L in HMSR to a maximum of 40.70 ± 16.64 mg/L in WSR. The mean BOD5 concentrations ranged from a minimum of 0.157 ± 0.124 mg/L in EER to a maximum of 2.390 ± 0.457 mg/L in MR. The spatial trends of BOD5 were similar to those of CODMn, with high values in HR and downstream (MR, TJR, and WSR) and low values in HMR and upstream (EER, LGR, and HMSR). The mean values of TDP ranged from a minimum of 0.056 ± 0.016 mg/L in LGR to a maximum of 0.086 ± 0.032 mg/L in MR. The mean DRP concentrations ranged from a minimum of 0.017 ± 0.006 mg/L in LGR to a maximum of 0.031 ± 0.008 mg/L in TJR. The mean values of NH4 +–N and NO3 −–N ranged from a minimum of 0.134 ± 0.151 and 0.063 ± 0.078 mg/L in HMSR to a maximum of 0.562 ± 0.249 and 0.563 ± 0.58 mg/L in TJR. NO2 −–N ranged from 0.004 ± 0.004 mg/L in MR to 0.012 ± 0.066 mg/L in TJR.

Fig. 4
figure 4

Spatial variations of DOC, POC, BOD5, CODMn, TDP, DRP, NH4 +–N, NO3 −–N and NO2 −–N, and Chl-a in different study regions. Note: the sample sites from upstream to downstream are EER, LGR, HMSR, HMR, HR, MR, TGR, and WSR, respectively

Full size image

Correlation among PARAFAC components and water quality parameters

CODMn and BOD5 were used as basic water quality indexes to estimate the organic matter in the Heilongjiang River. CODMn was significantly correlated with POC, and fluorescence component C3, and the presence of BOD5 was complemented by the increased POC and fluorescence component C3 (Fig. 5), which suggests that POC and fluorescence component C3 made a great contribution to the CODMn in the Heilongjiang River.

Fig. 5
figure 5

Correlation between BOD5, CODMn, and organic carbon (component C3, and POC). The unit of fluorescence intensity is calibrated and normalized in Raman Units (R.U.), and the sample numbers is 160. Note: the sample sites from upstream to downstream are EER, LGR, HMSR, HMR, HR, MR, TGR, and WSR, respectively

Full size image

As shown in Fig. 6, the fluorescence components were significantly correlated with the selected water quality parameters. Significant relationships were found between NH4 +–N, DRP, Chl-a, and component C1 (r = 0.273, −0.375, and 0.515; p < 0.001, p < 0.001, and p < 0.001; n = 160) (Fig. 6 a–c). Component C2 was highly correlated with DRP and Chl-a (r = 0.290 and 0.431; p < 0.001 and p < 0.001; n = 160) (Fig. 6d, e). Furthermore, significant linear relationships were found between POC, NO2 −–N, TDP, Chl-a, and fluorescence component C3 (r = 0.213, 0.200, 0.183, and −0.179; and p < 0.01, p < 0.05, p < 0.05, and p < 0.01; n = 160) (Fig. 5f–i), and component C4 was significantly correlated with NH4 +–N, DRP, and Chl-a (r = 0.289, −0.322, and 0.474; p < 0.001, p < 0.001, and p < 0.001; n = 160) (Fig. 6j–l).

Fig. 6
figure 6figure 6

Correlation among the four fluorescence components and general water quality properties. The unit of fluorescence intensity is calibrated and normalized in Raman Units (R.U.), and the sample number is 160. Note: the sample sites from upstream to downstream are EER, LGR, HMSR, HMR, HR, MR, TGR, and WSR, respectively

Full size image

Discussion

Characteristic of CDOM in the Heilongjiang River

In this study, all of the fluorescence components were assigned as humic-like material, but the spectral characteristics of the protein-like fluorophore were not detected in Heilongjiang (Table 2). The spectral characteristics of the excitation and emission of these four components presented in this study fit well with the components previously identified by other authors and reported in the literature (Table 2). In some other inland water studies such as at White Clay Creek (Cory and Kaplan, 2012), Xinan River (Liu et al. 2014), Taihu Lake (Zhang et al. 2014), and Yangtze River (Wang et al. 2015), protein-like components were detected, but in the Heilongjiang River, protein-like components were not detected. The absence of protein-like components has seldom been noted in other lake and river water studies. The protein-like component is strongly associated with bioavailable and liable organic substrates as well as degraded peptide material (Hudson et al., 2008). Compared with humic-like components, protein-like fluorescence components are more likely used by microbes as an energy source. In a previous study, the protein-like component was used as an indicator of organic pollution (Baker and Inverarity 2004; Hur and Cho, 2012) derived from anthropogenic activities. Hence, the absence of the protein-like fluorescence component in the Heilongjiang River indicates that the water quality of the Heilongjiang was not significantly influenced by anthropogenic activities.

Four components were identified by the PARAFAC modeling of the CDOM fluorescence spectroscopic data from the Heilongjiang River. The microbial humic-like component C1 generally had the highest relative distribution (approximately 50 %), higher than those of terrestrial humic-like components C2–4 in each sample site. This result is similar to that in Tainmuhu Lake (Zhang et al. 2011), suggesting that microbial humic-like organic matter was an important source for these four components in the Heilongjiang River. The F max of fluorescence components C1, C2, and C4 were almost perfectly linearly correlated (p < 0.001) (Fig. 3), indicating that they are controlled by the same factors (Liu et al. 2014). However, component C3 was not correlated with the other three components. This difference in the behavior of specific components from the general trend could be identified by the use of the PARAFAC technique, which helped identify the difference in C3 that was apparent in the single fluorescence excitation (Table 1). In addition, the total relative percentage (%) of terrestrial humic-like components (C2, C3, and C4) in each study site ranges from 47 to 58 %. This indicates that the concentration of fluorescence components at the sample site may be correlated to the terrestrial input in the Heilongjiang River. There are the Greater Khingan Mountains in the southwest of the upstream of Heilongjiang, the Lesser Khingan Mountains in the southwest of the midstream, and the Changbai mountains in the southwest of the downstream (Fig. 1). First, the inland tributaries would take partly terrestrial DOM into Heilongjiang. Second, the mountain runoff formed by the melting snow or rain carries a large amount of organic substances. As reported in the literature, the humic fraction was the largest soil organic carbon fraction of the black soil in the northeast region of China (Heilongjiang area) (Xing et al., 2005). The inland tributaries and mountains runoff are rich in organic matter. Therefore, the CDOM components in Heilongjiang are mainly influenced by the terrestrial input organic matter. The delivery of the terrestrial DOM subsidy plays an important role in catchment carbon budgets influencing the ecology of surface waters. In addition, the inputs of allochthonous DOM into the watershed are tied to the carbon cycle and influence the environmental aspects of the Heilongjiang River.

Relationships between BOD5, CODMn, and fluorescence components in the Heilongjiang River

The degree of organic pollution that occurs due to an excessive amount of organic matter has typically been monitored by measuring BOD5 and CODMn values in rivers (Hur and Cho, 2012). As previous studies reported, protein-like components make a great contribution to BOD5 (Baker and Inverarity, 2004). In the Heilongjiang River, there were significant positive relationships between terrestrial humic-like component C3 (r = 0.433, p < 0.001; n = 160), POC (r = 0.564, p < 0.001; n = 160), and BOD5 (Fig. 5). This result indicates that terrestrial humic-like component C3 may easily be used by microbes as a carbon source and the input of terrestrial organic matter would have an influence on the biomass in the Heilongjiang River. However, component C3 in the Heilongjiang River is not a major component (Fig. 2). These results indicate that the bioavailability of DOM in the Heilongjiang River is not high. The presence of CODMn was complemented by the increased POC (r = 0.398, p < 0.001; n = 160) and component C3 (r = 0.187, p < 0.05; n = 160) (Fig. 5), which suggests that POC and component C3 make great contributions to CODMn in the Heilongjiang River. POC is a labile C pool in the forest soil system, which enters the ecosystem mainly from litter fall and leaves the ecosystem via soil respiration, lateral transport, and seepage (Sun and Wang, 2014). Therefore, the significant positive relationships between BOD5, CODMn, and POC indicate that POC transported from the forest system into water make a contribution to the carbon cycle in the Heilongjiang River. Furthermore, the fluorescence of terrestrial humic-like component C3 shows significant positive relationships with BOD5 and CODMn. These results suggests that the water quality of the Heilongjiang River was mainly influenced (represented by CODMn and BOD5) by terrestrial inputs organic material. A high level of BOD5 deteriorates the river water quality by the rapid decomposition of biodegradable organic matter and the subsequent depletion of dissolved oxygen, while CODMn traditionally represents the total organic matter. In Heilongjiang, high levels of BOD5 and CODMn occur at HR and downstream (Fig. 4), which can be attributed to the terrain. Many mountains are seated beside the HR and its downstream watershed, such as LKM, CBM (Fig. 1). Melting snow or runoff from the mountains that may carry a large amount of DOM would influence the quality of the water. Although there is also a GKM upstream of HR in Heilongjiang, owing to the geographical position of GKM, its melting snow or runoff would not enter the Heilongjiang River in a great amount and produce a great impact on the water (Fig. 1).

Relationships between fluorescence components and water quality parameters

Eutrophication poses severe threats to the quality, safety, and sustainability of water resources worldwide (Heisler et al. 2008; Qin et al. 2010). Phosphorus and nitrogen indexes that could evaluate the level of eutrophication were also tested in terms of the relationships with fluorescence components (Fig. 6). The indexes of dissolved phosphorus and dissolved nitrogen can be attributed to the turnover of DOM (Qualls and Haines 1991; Hopkinson et al. 2002; Kragh and SØndergaard, 2004). Therefore, as an exporter of phosphorus and nitrogen, DOM plays an important role in terrestrial and stream ecosystems. Bacteria and algae in the water can use these nutrients for their growth. After algae die, their residues partly indirectly contribute to the content of DOM. Significant correlations between the fluorescence components and water quality parameters indicate a common controlling factor of the components and the selected parameters concentrations in the Heilongjiang River. In the Heilongjiang River, Chl-a showed significantly positive relationships with component C1 (r = 0.264, p < 0.001; n = 160), C2 (r = 0.239, p < 0.001; n = 160), and C4 (r = 0.263, p < 0.001; n = 160), but significantly negative relationships with component C3 (r = 0.224, p < 0.01; n = 160) (Fig. 6). Notably, the correlation coefficient between Chl-a and component C1 was higher than that between Chl-a and C2 or Chl-a and C4 (Fig. 6); this was partly attributed to higher relation of Chl-a and C1. The presence of C1 in our water samples was similar to microbial oxidized components (Cory and Mcknight, 2005) and phytoplankton degradation releases humic components due to microbial activity (Zhang et al., 2011). The appearance of component C1 in our river samples suggests a microbial humic component originating from the microbial transformation products. Previous studies have demonstrated that this component may be produced autochthonously in aqueous environments (Borisover et al. 2009, and Zhang et al. 2011). Chl-a is an important component in the cells of phytoplankton. After phytoplankton dying, the Chl-a would be released from the phytoplankton cells to the water, while some organic matter of phytoplankton residue would also be transformed into C1 (Cory and Mcknight, 2005; Zhang et al., 2011). Component C3 could be used as a carbon source for microbes, so some phosphorus and nitrogen would be released from the biodegradation of component C3. The released phosphorus and nitrogen could help phytoplankton growth. In addition, fluorescence components C1, C2, and C4 show significant negative relationships with DRP. The concentration of NH4 +–N generally had negative relationships with fluorescence components C1 and C4. Fluorescence component C3, which has a negative relationship with Chl-a, shows significant relationship with NO2 −–N and TDP, which may indicate that the dynamic of component C3 is different from those of components C1, C2, and C4 in the Heilongjiang River.

Conclusions

In this study, four fluorescence components C1 (microbial humic-like component), C2–4 (terrestrial humic-like) were identified in the Heilongjiang River. The result of statistical analysis shows that terrestrial humic-like component C3 is easily used by microbes as a carbon source in the Heilongjiang River, and POC may make a great contribution to CODMn. The results of BOD5 and CODMn in HR and its downstream showed high values, indicating that melting snow or runoff from the mountains that may carry large amounts of DOM would influence the quality of the water. Relevant analysis results showed that the phytoplankton residue would also be turned over into C1 by microbial transformation and C3 could be potential nutrients for phytoplankton growth in the Heilongjiang River.