Abstract

Dissolved organic carbon (DOC) and chromophoric dissolved organic matter (CDOM) in rivers and reservoirs on the western Loess Plateau, which is an area of severe soil erosion, were investigated in September 2017 to analyze the CDOM characteristics and composition, DOC distribution and influence of environmental factors on these parameters. Great differences of water parameters were exhibited between different groups based on the analysis of variance (p < 0.01). The results indicated that rivers exhibited higher DOC concentrations (mean: 3.70 mg/L) than reservoir waters (mean: 2.04 mg/L). Artificial and agricultural lands exert a large influence on DOC concentrations, which verifies the hypothesis that intense anthropogenic activity results in high DOC concentrations. The CDOM absorption at 350 nm [aCDOM(350)] of tributary water samples was 2.73 m−1, which was higher than that in the Yellow River (1.71 m−1) and reservoir waters (1.33 m−1). The effects of DOC, TC and turbulence (Tur) on CDOM are positive and significant (p < 0.05) according to the multiple linear regressions. An analysis of the optical characteristics of CDOM indicated that waters on the Loess Plateau contained abundant humic acid and higher levels of allochthonous DOM with a higher molecular weight (MW) based on the spectral slopes (S) and specific UV absorbance (SUVA254) values.

INTRODUCTION

Dissolved organic carbon (DOC) is the major fraction of dissolved organic matter (DOM) active in the global carbon cycle (Battin et al. 2008) and influences light penetration, the water chemical environment and the mobility and toxicity of heavy metals (Brooks et al. 2007). DOC mineralization caused by biodegradation or photodegradation is one of the major sources of CO2 in the atmosphere (Song et al. 2018). Therefore, a greater understanding of the DOC distribution and characteristics in limnological environments is of fundamental importance to the carbon cycle in aquatic environments.

Significant effort has been dedicated to examining chromophoric dissolved organic matter (CDOM) as a proxy by which to assess aquatic DOC information using remote sensing over various water bodies (Del Castillo & Miller 2008). The CDOM, the light-absorbing fraction of the DOM, plays an important role in biogeochemical processes with respect to the absorption of ultraviolet radiation, thereby protecting organisms from UV damage, reducing the amount and quality of photosynthetically active radiation available to phytoplankton and decreasing the primary productivity (Guéguen et al. 2005). There are two sources of CDOM: allochthonous, which originates primarily from vascular plants and soil organic matter of the catchment area, and autochthonous, which is derived from the decomposition of algae and aquatic vegetation within a water body (Graeber et al. 2012; Brezonik et al. 2015). Former studies have confirmed that aromatic, phenolic, and acidic functional groups in different sizes of molecules were found among CDOM molecules (Brezonik & Arnold 2011; Brezonik et al. 2015). Changes in the CDOM optical characteristics reflect the variations in CDOM composition, which result from physical, chemical and biological processes that occur in the water column (Coble 1996). The ratio of absorption at 250 and 365 nm (E250:365), the specific UV absorbance at 254 nm (SUVA254), and the spectral slopes (S) are usually used to trace the source, chemical composition, and molecular weight of CDOM and the ratio of fulvic acid to humic acid (Weishaar et al. 2003; Helms et al. 2008).

Relationships between DOC concentration and CDOM absorption values, and significant positive correlations have been observed for most inland waters that receive conservative amounts of terrestrially derived DOC, which mainly consists of CDOM (Spencer et al. 2012; Toming et al. 2016). Despite its great potential, studies have also reported a high degree of uncertainty in relating CDOM to DOC in many scenarios (Chen et al. 2004; Brezonik et al. 2015; Toming et al. 2016). A high degree of uncertainty is involved when using observations from multiple watersheds of different ecosystems (landcover, soil, or climate), particularly with regard to the vegetation type (Li et al. 2018a). Previous studies have documented variations in allochthonous DOC and CDOM characteristics in relation to landcover changes (Piirsoo et al. 2012), soil types (Li et al. 2018a), hydrological conditions (Arvola et al. 2010), air temperature (Tian et al. 2013), altitude (Sobek et al. 2007) and atmospheric carbon dioxide concentration (Freeman et al. 2004). Therefore, the types of land use and human activities surrounding water bodies significantly influence DOC and CDOM with regard to nonpoint-source pollution levels and load mitigation measures (Wilson & Xenopoulos 2008; Specchiulli et al. 2018). DOC-CDOM relationships vary widely, and the dominant controls acting on these processes have yet to be fully understood (Huang & Chen 2009). In addition, numerous studies have reported strong relationships between DOC/CDOM properties and water quality characteristics, and the chlorophyll a (Chla), total suspended matter, total nitrogen (TN), total phosphorus contents, salinity and trophic status exhibit profound effects on DOC or CDOM optical properties (Gonnelli et al. 2013; Wen et al. 2016; Specchiulli et al. 2018). However, the relationships vary in diverse aquatic ecosystems, particularly for inland rivers and lakes, because of different allochthonous and autochthonous sources.

The Loess Plateau is bordered by the Gobi Desert and the Inner Mongolian Plateau to the north and located in the upper and middle reaches of the Yellow River. Over 70% of the Loess Plateau is a gully hill-dominated region caused by massive soil erosion (Zhao et al. 2013). The Loess Plateau is one of the most severely eroded regions in the world, and it is rich in soil carbon storage (2.682 Pg C in 2008) (Feng et al. 2013). Moreover, the plateau region has discharged large quantities of sediment into the Yellow River, thereby influencing the carbon cycle in the aquatic environment and resulting in DOC concentrations and CDOM absorptions that are distinctly different compared with those in other regions. The present study was undertaken in an effort to determine the distribution of DOC and CDOM in western Loess Plateau waters, which are characterized by various physiographic and environmental conditions as well as different anthropogenic pressures. The first objective was to evaluate the spatial distributions of DOC and CDOM on the western Loess Plateau and determine the reasons behind the observed distributions. Considering various environmental influences, we hypothesized that human activities and allochthonous sources would best explain the distribution. Moreover, the other two objectives were to examine the compositions and sources of CDOM based on optical variables and to identify the usage of CDOM in retrieval of DOC through remote sensing based on the DOC-CDOM relationships in arid and semiarid regions at high altitudes.

MATERIALS AND METHODS

Study area

The Loess Plateau is located in the northern region of central China, and it has an area of approximately 62.5 × 104 km2 (Figure 1) with about 47.2 × 104 km2 of the Loess Plateau being erosion-prone area (Zhao et al. 2013). Soil erosion has resulted in the region having the most vulnerable ecological environment in China, which restricts the sustainable development of the regional economy and society. The western Loess Plateau is a semiarid zone with an average altitude between 1,300 and 3,000 m and the characteristics of it are jointly influenced by the three plateaus. The land use varies greatly and includes forest (zonal distribution along the mounts), grassland and agriculture (between Laji Mount and Daban Mount), and several urban and industrial centers (Figure 1 and Figure S1, Supplementary Material). As a result, the study region provides a uniquely diverse mix of local environments that allow researchers to assess the robustness of spectroscopic methods for the study of DOC-CDOM relationships and analysis of characteristics and trace sources.

Figure 1

Location of the Loess Plateau in China and the distribution of samples collected in September 2017.

Figure 1

Location of the Loess Plateau in China and the distribution of samples collected in September 2017.

Water samples and quality determination

Water sampling was conducted from September 4 to September 14, 2017 at 65 sampling sites (Figure 1 and Table 1). At each site, surface water samples were collected approximately 0.2 m below the water surface. Water samples were collected in acid-washed high-density polyethylene (HDPE) bottles and placed in thermoelectric coolers while in the field. Chemical and physical parameters, e.g., pH, dissolved oxygen (DO), water temperature (T), and electrical conductivity (EC), were determined via in situ sampling using portable analyzers (EXO1, YSI, U.S.) while altitude was measured with a GPS. An approximately 2.5 L water sample was collected at each site for the determination of water quality parameters, including Chla, total nitrogen (TN), turbidity (Tur), DOC and total carbon (TC) in the laboratory. Water samples were kept at 4 °C until analysis, which generally occurred within two days. Chla concentration was measured based on the Lichtenthaler equation (Lichtenthaler et al. 1996). DOC concentration was measured by a TOC analyzer (vario TOC select, Elementar). Details of CDOM absorption measurements can be found in Shao et al. (2017). CDOM is characterized using the absorption coefficient at 350 nm for comparison purposes in this study. S was calculated using a non-linear fit of an exponential function to the absorption spectrum in the intervals of 350–400 nm (S350-400) and 280–400 nm (S280-400) using the equation in Spencer et al. (2012). SUVA254 was calculated as follows (Weishaar et al. 2003):
formula
(1)
Table 1

Chemical, biological and optical parameters of the water samples

ParametersYR (n = 13)
TYR (n = 16)
RYR (n = 36)
Mean ± SDMin-MaxMean ± SDMin-MaxMean ± SDMin-Max
T (°C) 15.65 ± 1.90 12.57–18.28 13.46 ± 2.13 9.80–19.19 17.88 ± 1.64 15.50–20.85 
EC (μs/cm) 384.97 ± 72.52 295–6,154.9 396.18 ± 158.40 230.00–1,014.00 370.36 ± 29.43 330.80–411.90 
DO (mg/L) 7.91 ± 0.67 7–9.19 7.49 ± 0.56 6.50–8.59 8.67 ± 0.18 8.40–8.92 
TC (mg/L) 30.10 ± 10.90 14.00–42.89 22.96 ± 13.11 11.97–61.40 37.14 ± 0.39 36.50–38.17 
TN (mg/L) 0.22 ± 0.21 0–0.47 1.54 ± 5.32 0–28.52 — — 
Altitude (m) 1,733.83 ± 271.57 1,478–2,355 2,187.92 ± 398.41 1,557–3,320 1,895.44 ± 215.41 1,733–2,172 
pH 8.02 ± 0.17 7.53–8.67 8.17 ± 0.27 7.79–8.86 8.28 ± 0.05 8.23–8.36 
Tur (NTU) 84.87 ± 100.66 5.46–405.21 153.28 ± 138.47 6.00–536.80 4.23 ± 0.37 3.72–4.96 
DOC (mg/L) 2.98 ± 2.26 0.59–7.05 4.42 ± 1.44 0.81–6.61 2.04 ± 0.52 1.27–3.40 
aCDOM (350)(m−11.71 ± 0.43 0.92–5.86 2.73 ± 1.35 0.81–6.57 1.33 ± 0.09 1.15–1.51 
SUVA254 (L mg−1 C−1 m−14.44 ± 2.97 1.30–9.20 3.73 ± 2.95 0.95–15.02 4.14 ± 0.92 2.45–6.17 
S350−400 (nm−115.78 ± 2.12 13.67–19.70 17.47 ± 4.03 11.65–27.56 12.21 ± 0.46 11.11–13.01 
S280−400 (nm−117.31 ± 1.67 14.93–19.48 18.01 ± 3.25 12.07–27.72 14.70 ± 0.78 13.73–17.25 
ParametersYR (n = 13)
TYR (n = 16)
RYR (n = 36)
Mean ± SDMin-MaxMean ± SDMin-MaxMean ± SDMin-Max
T (°C) 15.65 ± 1.90 12.57–18.28 13.46 ± 2.13 9.80–19.19 17.88 ± 1.64 15.50–20.85 
EC (μs/cm) 384.97 ± 72.52 295–6,154.9 396.18 ± 158.40 230.00–1,014.00 370.36 ± 29.43 330.80–411.90 
DO (mg/L) 7.91 ± 0.67 7–9.19 7.49 ± 0.56 6.50–8.59 8.67 ± 0.18 8.40–8.92 
TC (mg/L) 30.10 ± 10.90 14.00–42.89 22.96 ± 13.11 11.97–61.40 37.14 ± 0.39 36.50–38.17 
TN (mg/L) 0.22 ± 0.21 0–0.47 1.54 ± 5.32 0–28.52 — — 
Altitude (m) 1,733.83 ± 271.57 1,478–2,355 2,187.92 ± 398.41 1,557–3,320 1,895.44 ± 215.41 1,733–2,172 
pH 8.02 ± 0.17 7.53–8.67 8.17 ± 0.27 7.79–8.86 8.28 ± 0.05 8.23–8.36 
Tur (NTU) 84.87 ± 100.66 5.46–405.21 153.28 ± 138.47 6.00–536.80 4.23 ± 0.37 3.72–4.96 
DOC (mg/L) 2.98 ± 2.26 0.59–7.05 4.42 ± 1.44 0.81–6.61 2.04 ± 0.52 1.27–3.40 
aCDOM (350)(m−11.71 ± 0.43 0.92–5.86 2.73 ± 1.35 0.81–6.57 1.33 ± 0.09 1.15–1.51 
SUVA254 (L mg−1 C−1 m−14.44 ± 2.97 1.30–9.20 3.73 ± 2.95 0.95–15.02 4.14 ± 0.92 2.45–6.17 
S350−400 (nm−115.78 ± 2.12 13.67–19.70 17.47 ± 4.03 11.65–27.56 12.21 ± 0.46 11.11–13.01 
S280−400 (nm−117.31 ± 1.67 14.93–19.48 18.01 ± 3.25 12.07–27.72 14.70 ± 0.78 13.73–17.25 

Statistical analysis

The Pearson correlation coefficient (r), analysis of variance (ANOVA), hierarchical clustering analysis and multiple linear regressions (MLR) were calculated using SPSS 19.0. Principal component analysis (PCA) was conducted based on CANOCO 5.0. Riparian buffers of samples collected from rivers were computed based on ArcGIS 10.4 and the buffer width was determined according to Li et al. (2018b).

RESULTS

Water quality parameters

Hierarchical clustering analysis was used to distinguish the differences between samples collected on the western Loess Plateau. Two main groups can be distinguished from the dendrogram produced by this method (Figure S2, Supplementary Material). The first group contains 16 assorted samples collected in reservoirs (RYR) and most of the samples from the stem of the Yellow River (YR), while the second group includes most of the samples from the tributaries of the YR (TYR) and few samples from YR. Samples from RYR are easily distinguished from those collected in YR in the first group, as they are characterized by low concentrations of DOC and CDOM. As a result, samples were divided into three groups, YR, RYR and TYR.

The groups of samples exhibited large variations in water quality (Table 1). Significant differences of DOC and aCDOM(350) was also observed between these three groups (ANOVA, p < 0.01, Table s1). Moreover, the Tur of different groups exhibited a very notable variation (ANOVA, p < 0.01). The mean Tur in the TYR was higher than that in the YR, while RYR water samples exhibited the lowest Tur values (Table 1). However, there was no significant difference of EC in the groups (ANOVA, p > 0.5). Spatially, the mean DO concentration in the YR was higher than TYR and lower than RYR (8.67 ± 0.18 mg/L). The mean TC concentration the RYR was the highest and followed by samples in the TYR. However, TYR exhibited a higher concentration of TN (0.22 ± 0.21 mg/L) than that in YR (1.54 ± 5.32 mg/L).

PCA was performed with 10 water environment variables and 1 computed parameter for all the sampling sites (Figure 2). PC1 explained 37.04% and PC2 explained 17.58% of the variability among all the selected variables. The first PCA axis revealed gradients of C constituents. The DOC, Tur and aCDOM(350) contributed relatively high loadings in PC1, whereas TC showed high negative loadings. TN and EC showed positive loadings in PC2. A clear difference was found between river and reservoir waters (Figure 2). Samples collected from reservoirs clustered in close proximity to each other and were distributed on the negative side of PC1. Most samples in the TYR were clustered on the positive side of PC1 except several samples on the positive side of PC2 (Figure 2). YR samples clustered on the negative side of PC1, except samples located near Lanzhou City.

Figure 2

PCA of the physicochemical characteristics and loading data of factors of all waters collected in different rivers and reservoirs.

Figure 2

PCA of the physicochemical characteristics and loading data of factors of all waters collected in different rivers and reservoirs.

Figure 3

Spatial distribution of the DOC concentration of samples collected on the western Loess Plateau.

Figure 3

Spatial distribution of the DOC concentration of samples collected on the western Loess Plateau.

DOC in rivers and reservoirs on the Loess Plateau

The DOC concentrations significantly fluctuated and ranged from 0.59 to 7.05 mg/L, with a mean value of 3.15 mg/L (Table 1 and Figure 2). The river samples showed higher values than did the RYR, although the values of the rivers were less stable than those in RYR. Moreover, the mean DOC of TYR (4.42 mg/L) was higher than that of the YR (2.98 mg/L). High DOC concentrations were generally distributed in areas with larger populations (Figures 1 and 3). Two areas with high DOC concentrations were observed in the study area, namely, regions A (mean value: 4.83 mg/L) and B (mean value: 4.96 mg/L), as shown in Figure 2. The high DOC concentration areas were centered in Lanzhou and Xining Cities, which are the provincial capitals of Gansu and Qinghai Provinces, respectively. The highest DOC concentration was measured in sample s09 (7.05 mg/L), which was collected downstream of Lanzhou City, and the second highest concentration was measured in sample s52 (6.61 mg/L) near Xining City (Figures 1 and 2).

Moreover, the land uses in the riparian/buffer zone, which are thought to have a disproportionate impact on the transfer of water chemical materials, sediments and woody debris (Ye et al. 2014), were imported to analyze how the DOC concentration was affected by the land use type (except RYR, Figure S3, Supplementary Material, and Table 2). Riparian zones dominated by agriculture exhibited higher DOC levels than those of other land types, followed by artificial areas. However, the DOC range of riparian zones dominated by artificial areas varied more greatly than did that of agricultural land (Table 2).

Table 2

Statistics of the DOC, aCDOM(350) and DOC-to-CDOM ratio in different riparian buffer zones dominated by artificial areas, agricultural land and grassland

Land use typeParametersStatistical parameters
Min-MaxMean ± SDCV
Artificial area DOC (mg/L) 1.18–7.05 3.80 ± 2.23 58.71% 
aCDOM(350) 1.15–2.76 2.12 ± 0.52 24.77% 
Agricultural land DOC (mg/L) 1.53–6.61 4.65 ± 1.30 27.91% 
aCDOM(350) 0.81–6.57 2.81 ± 1.50 53.49% 
Grass land DOC (mg/L) 1.59–5.19 2.49 ± 1.93 77.65% 
aCDOM(350) 1.42–2.75 1.85 ± 0.44 23.85% 
Land use typeParametersStatistical parameters
Min-MaxMean ± SDCV
Artificial area DOC (mg/L) 1.18–7.05 3.80 ± 2.23 58.71% 
aCDOM(350) 1.15–2.76 2.12 ± 0.52 24.77% 
Agricultural land DOC (mg/L) 1.53–6.61 4.65 ± 1.30 27.91% 
aCDOM(350) 0.81–6.57 2.81 ± 1.50 53.49% 
Grass land DOC (mg/L) 1.59–5.19 2.49 ± 1.93 77.65% 
aCDOM(350) 1.42–2.75 1.85 ± 0.44 23.85% 

Note: the units of the DOC and CDOM are mg/L and m−1, respectively.

CDOM spectral characteristics

Clearly, the aCDOM(350) of TYR (2.73 ± 1.35 m−1) was higher than those of YR (1.71 ± 0.43 m−1) and RYR (1.33 ± 0.09 m−1) (Table 1). The highest values were obtained in sample s52 (6.57 m−1) followed by s54 (6.03 m−1) and s53 (5.88 m−1). The samples with the three highest aCDOM(350) values were all located in Huangzhong County next to Xining City. Moreover, the aCDOM(350) values of the RYR samples were all below 1.52 m−1 (Table 1), but the highest value was measured in sample s22 (1.51 m−1) collected in the mouth of the Daxia River, which flows into the Liujiaxia Reservoir. In addition, 89.48% of the TYR samples showed high CDOM concentrations (higher than 1.5 m−1), with 50% of the samples producing absorption values exceeding 2.5 m−1, and these values were higher than those in the YR and RYR waters. The samples with the four highest measured values in the study were all collected from the TYR upstream of the Huangshui River near Xining City, and they were located geographically close to each other.

The measured S values of the RYR samples were higher than the YR values, while the TYR samples exhibited the lowest S values. Moreover, S exhibited significant variation in the YR and TYR samples but remained relatively stable in the RYR samples. The SUVA254 values in the TYR water samples ranged from 0.95 to 15.02 L mg−1 C−1 m−1, while the values in the YR water samples varied from 1.30–9.20 L mg−1 C−1 m−1. The lowest SUVA254 value in the RYR water samples was 2.45 L mg−1 C−1 m−1, and the highest value was 6.17 L mg−1 C−1 m−1. The mean SUVA254 value of the YR samples (4.63 ± 2.80 L mg−1 C−1 m−1) was clearly higher than those of the TYR and RYR samples (Table 1).

DOC and CDOM vs. water parameters

The DOC was negatively correlated with the DO (Table 3 and Figure S4, Supplementary Material), and a strong correlation was also obtained between DOC and TC in western Loess Plateau waters. However, altitude and pH correlated weakly with DOC for all water samples. A positive correlation was obtained between DOC and Tur (Figure S4e). Almost no correlations were observed between DOC and CDOM in YR, RYR and TYR, while a positive correlation was obtained for all water samples in this study (Table 3 and Figure S4f).

Table 3

Correlations between the DOC and water quality parameters for all the samples

ParametersWater temperature (°C)DO (mg/L)aCDOM(350) (m−1)TC (mg/L)Altitude (m)pHTur (NTU)
Correlation with DOC −0.50** −0.64** 0.47** −0.76** 0.19 −0.33** 0.53** 
ParametersWater temperature (°C)DO (mg/L)aCDOM(350) (m−1)TC (mg/L)Altitude (m)pHTur (NTU)
Correlation with DOC −0.50** −0.64** 0.47** −0.76** 0.19 −0.33** 0.53** 

**Correlation is significant at the 0.01 level (2-tailed).

The CDOM concentration varied greatly among the different water bodies and exhibited large differences in its correlations with other quality parameters (Tables 2 and 4). Significant correlations were obtained between the EC and aCDOM(350) in the YR and RYR samples. However, almost no correlations were observed in the TYR samples. Tur was positively correlated with aCDOM(350) for river samples (YR and TYR), while no correlation was obtained in the RYR samples. Positive but not significant correlations were observed between the TN concentration and aCDOM(350) in the YR, TYR and all samples. Moreover, DOC and altitude exhibited significant correlations with aCDOM(350) in RYR, but a negative correlation was observed between DO and CDOM for all the samples (Table 4).

Table 4

Correlations between the CDOM and water quality parameters for all the samples and samples from YR, RYR and TYR, respectively

Correlation with CDOMYR (n = 13)RYR (n = 16)TYR (n = 36)All (n = 65)
Water temperature (°C) 0.42 0.57* −0.30 −0.48** 
EC (μs/cm) 0.82* 0.66** 0.13 0.18 
DO (mg/L) −0.43 0.59* −0.043 −0.39** 
DOC (mg/L) −0.16 −0.21 0.29 0.47** 
TC (mg/L) 0.34 −0.25 −0.086 −0.28* 
TN (mg/L) 0.88 —— 0.47 0.47 
Altitude (m) −0.47 −0.76** 0.25 0.37** 
pH 0.43 −0.63** −0.25 −0.21 
Tur (NTU) 0.54 0.084 0.53** 0.64** 
Correlation with CDOMYR (n = 13)RYR (n = 16)TYR (n = 36)All (n = 65)
Water temperature (°C) 0.42 0.57* −0.30 −0.48** 
EC (μs/cm) 0.82* 0.66** 0.13 0.18 
DO (mg/L) −0.43 0.59* −0.043 −0.39** 
DOC (mg/L) −0.16 −0.21 0.29 0.47** 
TC (mg/L) 0.34 −0.25 −0.086 −0.28* 
TN (mg/L) 0.88 —— 0.47 0.47 
Altitude (m) −0.47 −0.76** 0.25 0.37** 
pH 0.43 −0.63** −0.25 −0.21 
Tur (NTU) 0.54 0.084 0.53** 0.64** 

**Correlation is significant at the 0.01 level (2-tailed); *Correlation is significant at the 0.05 level (2-tailed).

MLR was used to investigate the impact of the above factors that significantly correlated with CDOM for all samples (T, DO, DOC, TC and Tur) (Table 5). Before the MLR, normalization and collinearity test are conducted at first. In the basic regression model, Tur significantly positively correlated with CDOM (p < 0.01). The effects of DOC and TC on CDOM are also positive and significant (p < 0.05), and the effects of T and DO are not significant.

Table 5

Regression results of several water parameters and CDOM for all the samples located on the Loess Plateau

ParameterCoefficientSig.VIFAdjusted R-squared
(constant) 0.070 0.895 — 0.49 
−0.472 0.151 1.92 
DO 0.171 0.694 1.69 
DOC 0.462** 0.015 3.33 
TC 0.515** 0.017 3.34 
Tur 0.253*** 0.000 1.50 
ParameterCoefficientSig.VIFAdjusted R-squared
(constant) 0.070 0.895 — 0.49 
−0.472 0.151 1.92 
DO 0.171 0.694 1.69 
DOC 0.462** 0.015 3.33 
TC 0.515** 0.017 3.34 
Tur 0.253*** 0.000 1.50 

** Significant at 0.05 level; *** at 0.01 level.

DISCUSSION

Factors affecting the DOC

No correlation was obtained between DOC and altitude (Table 4), which was inconsistent with the results of a previous study based on a large database of 7,514 lakes on six continents (Sobek et al. 2007). Generally, altitude inhibits human activity, increases photochemical degradation, decreases terrestrial input and presents combined climatic, topographic and edaphic effects, thereby resulting in negative relationships between DOC and altitude (Song et al. 2019). However, different relationships were observed between rivers and lakes due to their liquidity and complicated surrounding environments. The main reason may be that the number of samples was not high enough to support the correlation. The landscape, topography, precipitation, catchment characteristics and other environmental factors might explain the observed distributions (Sobek et al. 2007).

The results obtained in this study of high DOC concentrations in agricultural riparian buffers are consistent with the results reported by Graeber et al. (2012) and are in contrast to the results from previous studies in Canada and the USA (Wilson & Xenopoulos 2008). In our study, crop-based agriculture is the main type of agricultural area along the YR and tributaries; therefore, human disturbances (e.g. agricultural cultivation) aggravated the loss of soil OM and resulted in increasing terrestrial input into the rivers (Lambert et al. 2017). The other human-altered land use type, artificial areas, also exhibited high DOC concentrations (Table 2). Urbanization will generate possibilities for accelerated urban population expansion (Bai et al. 2014), and population is significantly correlated with the wastewater effluent (r = 0.95, p < 0.001) in China (Figure S5, Supplementary Material), consequently resulting in increasing DOC (nearly uncolored) levels from human sources (Brezonik et al. 2015; Griffin et al. 2018). Since the 1980s, Lanzhou and Xining Cities have undergone great urbanization (Figure S6, Supplementary Material). The highest DOC concentration (sample s09, downstream of Lanzhou City) might have provided an evidence for the interaction between populations and DOC concentrations (Figure 2). The DOC in rivers increases with the population density and the proportion of human-dominated landscapes (Butman et al. 2015); thus, land use/cover changes affected by humans in sub-catchments might be a key factor affecting river DOC concentrations and characteristics in this study, confirming the hypothesis that human activities influence the DOC distribution, consequently.

CDOM distribution and sources

The result that aCDOM(350) was high in the riparian buffer zones dominated by agriculture and artificial areas demonstrated that land use affected intensely by human activities brought great influences to the CDOM concentration. Moreover, waters on the western Loess Plateau were influenced by the surrounding environments because of the soil erosion and low vegetation cover (Meng et al. 2008). Results obtained from the MLR of positive and significant effects of Tur on CDOM showed that different allochthonous sources from the catchment area might be the main reason for the distribution of CDOM, which also confirmed the hypothesis.

The mean S350-400 values of the YR and TYR water samples were higher than those of the Yangtze River (Chen et al. 2013) and the Elizabeth River (0.0149 nm−1) (Helms et al. 2008) but were lower than those in the Yukon River (Spencer et al. 2009). The mean S280-400 values were higher than those in the Suwannee River (Xiao et al. 2013) but lower than those in the Liaohe River (Shao et al. 2017) and terminal lakes (Zhang et al. 2005). Former studies have demonstrated that a steeper S typically corresponds to decreasing molecular weight and aromaticity of the CDOM. Consequently, the CDOM MW in the rivers on the western Loess Plateau was higher than those in the Yangtze, Suwannee and Elizabeth Rivers, while it was lower than those of rivers located at high latitudes. Moreover, CDOM MW in the TYR samples was lower than those in the YR and RYR samples for high S values (Table 1), thus confirming the argument that the long residence time of reservoirs combined with long sunshine duration might cause enhanced photochemical degradation and result in a lower CDOM MW (Twardowski & Donaghay 2002; Zhang et al. 2009).

Previous studies have confirmed that SUVA254 is positively correlated with both the degree of aromaticity and MW (Weishaar et al. 2003; Cawley et al. 2012). Thus, we inferred that the aromatic moieties of DOM in the RYR water samples were lower compared with those of river samples (YR and TYR) for lower SUVA254 values. High SUVA254 values are reported to be associated with allochthonous-dominated sources, whereas low SUVA254 values are considered to be related to microbially-dominated substances in DOC (Weishaar et al. 2003; Spencer et al. 2012). Photochemical degradation and biological activities could result in decomposition of the aromatic groups of DOM, and these degradation effects may become enhanced with prolonged water residence times in reservoir waters, thus resulting in the lower MW aromatic moieties of CDOM in the reservoir water samples. The water samples in this study exhibited significantly higher SUVA254 values (mean value: 4.10 L mg−1 C−1 m−1) than those of rivers and lakes, which were greatly influenced by autochthonous-dominated DOM (Spencer et al. 2009, 2012). As a result, the SUVA254 values obtained in the study indicated that the CDOM composition in Loess Plateau waters was dominated by high MW fractions (Weishaar et al. 2003), which is in agreement with the result obtained from the S values.

Relationship between the DOC and CDOM

No correlation was obtained between DOC and CDOM in the YR, RYR and TYR samples, which was inconsistent with previous studies that found strong positive correlations between the two parameters in a wide range of aquatic systems (Zhang et al. 2005; Spencer et al. 2012; Zhao et al. 2016). Linear relationships between the DOC and CDOM have usually been observed during a short period or a single season without significant temperature variations (Li et al. 2018a), while decoupled relationships between the two parameters were mostly based on observations across watersheds/rivers and seasons (Tian et al. 2013). Previous studies focusing on the Yangtze and Congo Rivers and waters across mainland USA also reported that the relationship between CDOM and DOC was not conservative due to estuarine mixing or photodegradation (Chen et al. 2004; Spencer et al. 2009, 2012). The spatiotemporal differences in DOC-CDOM correlations may have been largely due to different DOC and CDOM sources (Mitrovic & Baldwin 2016). As Chen et al. (2004) confirmed, only if the CDOM and uncolored dissolved organic matter (UDOM) ratio was constant would there be a strong relationship between CDOM and DOC. For the RYR samples, a larger quantity of UDOM generated by photo- or biodegradation might have resulted in a weak correlation. In comparison, the DOC concentrations in the Lijiaxia Reservoir were higher, but the CDOM level was lower than those of the samples from the Liujiaxia Reservoir (Table 1). Unsynchronized DOC and CDOM variations in samples from the two reservoirs resulted in poor relationships in the RYR samples (Table 4). The main stem of the Huangshui River passes through Xining and Lanzhou Cities and carries large quantities of industrial wastewater and sewage that contains larger loads of uncolored DOC (Brezonik et al. 2015), which might have significantly influenced the DOC and CDOM relationship in the TYR samples (r = 0.29). Although the Datong River is a tributary of the Huanghsui River, the main land use types surrounding the stem of the Huangshui River were agricultural and grassland, but the Datong River was surrounded by forestland (Figure S2). When the Datong River samples were eliminated from the data set, an improved correlation (r = 0.610) was exhibited between DOC and CDOM levels. Moreover, the DOC-CDOM relationship weakened with increasing wavelengths due to the inaccuracy of CDOM absorption measurements at longer wavelength (Spencer et al. 2012). If the CDOM absorption at 254 nm was used, improved results were obtained, although they were still inferior to those reported in other studies that focused on rivers in plains and arctic areas (Ågren et al. 2007; Shao et al. 2016). As a result, using CDOM alone as a predictor of DOC concentration is subject to a substantial degree of uncertainty, particularly in optically complex waters (Toming et al. 2016). Unfortunately, sub-basins were not extracted for the region studied in this paper because it did not represent an integrated area. Consequently, basin characteristics affected by DOC-CDOM could not be identified due to the aforementioned limitations. Therefore, we argue that additional data should be collected and mined in great detail to better understand the reasons behind the weak relationship between DOC and CDOM on the western Loess Plateau.

CONCLUSIONS

The Loess Plateau is a complicated and unique area that is subject to severe soil erosion and hosts an abundant amount of soil carbon storage. A preliminary study was conducted to provide an overall view of the DOC distribution and CDOM absorption characteristics and insights into the factors influencing the correlation between DOC and CDOM on the western Loess Plateau. The DOC concentrations significantly fluctuated in the tributary water samples, while those of the reservoir water samples remained relatively stable. Higher DOC concentrations were mainly distributed in artificial and agricultural lands, which verified the hypothesis. Positive correlations were obtained between DOC and TC while weak correlations were observed between DOC and aCDOM(350) in the waters of the western Loess Plateau. The values of aCDOM(350) in the TYR were higher than those in the YR and RYR, which was inconsistent with the DOC distribution results. The effects of DOC, TC and Tur on CDOM are positive and significant (p < 0.05). However, samples that had been collected in riparian buffers dominated by crop land and built-up areas – particularly crop land – exhibited high CDOM concentrations. The CDOM composition in Loess Plateau water bodies was dominated by high MW fractions based on measured SUVA254 and S values. The study area was not an integrated watershed, and no sub-watersheds were partitioned to incorporate the basin characteristics that affect DOC and CDOM levels, so further research should be conducted to quantify the associated contributions. Moreover, seasonal variations in the relationships between DOC and CDOM may occur for waters on the Loess Plateau, which is a subject that will be studied further in the future.

ACKNOWLEDGEMENTS

This research was jointly supported by the National Natural Science Foundation of China (No. 41601377) and the Soft Science Research of Technology Development Projects of Henan Province (No. 192400410083).

SUPPLEMENTARY MATERIAL

The Supplementary Material for this paper is available online at https://dx.doi.org/10.2166/wst.2020.004.

REFERENCES

Ågren
A.
Buffam
I.
Jansson
M.
Laudon
H.
2007
Importance of seasonality and small streams for the landscape regulation of dissolved organic carbon export
.
Journal of Geophysical Research: Biogeosciences
112
(
G3
),
1
11
.
Arvola
L.
Rask
M.
Ruuhijärvi
J.
Tulonen
T.
Vuorenmaa
J.
Ruoho-Airola
T.
Tulonen
J.
2010
Long-term patterns in pH and colour in small acidic boreal lakes of varying hydrological and landscape settings
.
Biogeochemistry
101
(
1–3
),
269
279
.
Bai
X. M.
Shi
P. J.
Liu
Y. S.
2014
Realizing China's urban dream
.
Nature
509
,
158
160
.
Battin
T. J.
Kaplan
L. A.
Findlay
S.
Hopkinson
C. S.
Marti
E.
Packman
A. I.
Newbold
J. D.
Sabater
F.
2008
Biophysical controls on organic carbon fluxes in fluvial networks
.
Nature Geoscience
1
(
2
),
95
.
Brezonik
P. L.
Arnold
W. A.
2011
Water chemistry: An introduction to the chemistry of natural and engineered aquatic systems
.
Oxford University Press
,
New York, NY
(
782
pp.).
Brezonik
P. L.
Olmanson
L. G.
Finlay
J. C.
Bauer
M. E.
2015
Factors affecting the measurement of CDOM by remote sensing of optically complex inland waters
.
Remote Sensing of Environment
157
,
199
215
.
Butman
D. E.
Wilson
H. F.
Barnes
R. T.
Xenopoulos
M. A.
Raymond
P. A.
2014
Increased mobilization of aged carbon to rivers by human disturbance
.
Nature Geoscience
8
(
2
),
112
116
.
Chen
H.
Meng
W.
Zheng
B.
Wang
C. Y.
An
L. H.
2013
Optical signatures of dissolved organic matter in the watershed of a globally large river (Yangtze River, China)
.
Limnologica – Ecology and Management of Inland Waters
43
(
6
),
482
491
.
Freeman
C.
Fenner
N.
Ostle
N. J.
Kang
H.
Dowrick
D. J.
Reynolds
B.
Sleep
D.
Hughes
S.
Hudson
J.
2004
Export of dissolved organic carbon from peatlands under elevated carbon dioxide levels
.
Nature
430
(
6996
),
195
.
Graeber
D.
Gelbrecht
J.
Pusch
M. T.
Anlanger
C.
von Schiller
D.
2012
Agriculture has changed the amount and composition of dissolved organic matter in Central European headwater streams
.
Science of the Total Environment
438
,
435
446
.
Griffin
C. G.
Finlay
J. C.
Brezonik
P. L.
Olmanson
L.
Hozalski
R. M.
2018
Limitations on using CDOM as a proxy for DOC in temperate lakes
.
Water Research
144
,
719
727
.
Guéguen
C.
Guo
L.
Tanaka
N.
2005
Distributions and characteristics of colored dissolved organic matter in the western Arctic Ocean
.
Continental Shelf Research
25
(
10
),
1195
1207
.
Helms
J. R.
Stubbins
A.
Ritchie
J. D.
Minor
E. C.
Kieber
D. J.
Mopper
K.
2008
Absorption spectral slopes and slope ratios as indicators of molecular weight, source, and photobleaching of chromophoric dissolved organic matter
.
Limnology and Oceanography
53
(
3
),
955
969
.
Huang
W.
Chen
R. F.
2009
Sources and transformations of chromophoric dissolved organic matter in the Neponset River Watershed
.
Journal of Geophysical Research: Biogeosciences
114
(
G4
),
1
14
.
Lambert
T.
Bouillon
S.
Darchambeau
F.
Morana
C.
Roland
F. A.
Descy
J. P.
Borges
A. V.
2017
Effects of human land use on the terrestrial and aquatic sources of fluvial organic matter in a temperate river basin (The Meuse River, Belgium)
.
Biogeochemistry
136
(
2
),
191
211
.
Li
J.
Yu
Q.
Tian
Y. Q.
Boutt
D. F.
2018a
Effects of landcover, soil property, and temperature on covariations of DOC and CDOM in inland waters
.
Journal of Geophysical Research: Biogeosciences
123
(
4
),
1352
1365
.
Meng
Q.
Fu
B.
Tang
X.
Ren
H.
2008
Effects of land use on phosphorus loss in the hilly area of the Loess Plateau, China
.
Environmental Monitoring and Assessment
139
(
1–3
),
195
204
.
Piirsoo
K.
Viik
M.
Kõiv
T.
Käiro
K.
Laas
A.
Nõges
T.
Pall
P.
Selberg
A.
Toomsalu
L.
Vilbaste
S.
2012
Characteristics of dissolved organic matter in the inflows and in the outflow of Lake Võrtsjärv, Estonia
.
Journal of Hydrology
475
,
306
313
.
Shao
T. T.
Song
K.
Jacinthe
P. A.
Du
J.
Zhao
Y.
Ding
Z.
Guan
Y.
Zhang
B.
2016
Characteristics and sources analysis of riverine chromophoric dissolved organic matter in Liaohe River, China
.
Water Science and Technology
74
(
12
),
2843
2859
.
Shao
T. T.
Zheng
H.
Song
K. S.
Zhao
Y.
Zhang
B.
2017
Influence of environmental factors on absorption characteristics of suspended particulate matter and CDOM in Liaohe River watershed, northeast China
.
Environmental Science and Pollution Research
24
(
23
),
19322
19337
.
Sobek
S.
Tranvik
L. J.
Prairie
Y. T.
Kortelainen
P.
Cole
J. J.
2007
Patterns and regulation of dissolved organic carbon: an analysis of 7,500 widely distributed lakes
.
Limnology and Oceanography
52
(
3
),
1208
1219
.
Song
K.
Wen
Z.
Xu
Y.
Yang
H.
Lyu
L.
Zhao
Y.
Fang
C.
Shang
Y. X.
Du
J.
2018
Dissolved carbon in a large variety of lakes across five limnetic regions in China
.
Journal of Hydrology
563
,
143
154
.
Song
K.
Shang
Y.
Wen
Z.
Jacinthe
P. A.
Liu
G.
Lyu
L.
Fang
C.
2019
Characterization of CDOM in saline and freshwater lakes across China using spectroscopic analysis
.
Water Research
150
,
403
417
.
Specchiulli
A.
Cilenti
L.
D'Adamo
R.
Fabbrocini
A.
Guo
W.
Huang
L.
Padedda
B. M.
Scirocco
T.
Magni
P.
2018
Dissolved organic matter dynamics in Mediterranean lagoons: the relationship between DOC and CDOM
.
Marine Chemistry
202
,
37
48
.
Spencer
R. G. M.
Stubbins
A.
Hernes
P. J.
Baker
A.
Mopper
K.
Aufdenkampe
A. K.
Dyda
R. Y.
Mwamba
V. L.
Mangangu
A. M.
Wabakanghanzi
J. N.
Six
J.
2009
Photochemical degradation of dissolved organic matter and dissolved lignin phenols from the Congo River
.
Journal of Geophysical Research: Biogeosciences
114
(
G3
),
1
12
.
Spencer
R. G. M.
Butler
K. D.
Aiken
G. R.
2012
Dissolved organic carbon and chromophoric dissolved organic matter properties of rivers in the USA
.
Journal of Geophysical Research: Biogeosciences
117
(
G3
),
1
14
.
Toming
K.
Kutser
T.
Tuvikene
L.
Viik
M.
Nõges
T.
2016
Dissolved organic carbon and its potential predictors in eutrophic lakes
.
Water Research
102
,
32
40
.
Twardowski
M. S.
Donaghay
P. L.
2002
Photobleaching of aquatic dissolved materials: absorption removal, spectral alteration, and their interrelationship
.
Journal of Geophysical Research: Oceans
107
(
C8
),
6-1
6-12
.
Weishaar
J. L.
Aiken
G. R.
Bergamaschi
B. A.
Fram
M. S.
Fujii
R.
Mopper
K.
2003
Evaluation of specific ultraviolet absorbance as an indicator of the chemical composition and reactivity of dissolved organic carbon
.
Environmental Science & Technology
37
(
20
),
4702
4708
.
Wen
Z. D.
Song
K. S.
Zhao
Y.
Du
J.
Ma
J. H.
2016
Influence of environmental factors on spectral characteristics of chromophoric dissolved organic matter (CDOM) in Inner Mongolia Plateau
.
China. Hydrology and Earth System Sciences
20
(
2
),
787
801
.
Xiao
Y. H.
Sara-Aho
T.
Hartikainen
H.
Vähätalo
A. V.
2013
Contribution of ferric iron to light absorption by chromophoric dissolved organic matter
.
Limnology and Oceanography
58
(
2
),
653
662
.
Ye
Y.
He
X.
Chen
W.
Yao
J.
Yu
S.
Jia
L.
2014
Seasonal water quality upstream of Dahuofang Reservoir, China–the effects of land use type at various spatial scales. CLEAN–Soil, Air
.
Water
42
(
10
),
1423
1432
.
Zhao
G.
Mu
X.
Wen
Z.
Wang
F.
Gao
P.
2013
Soil erosion, conservation, and eco-environment changes in the Loess Plateau of China
.
Land Degradation & Development
24
(
5
),
499
510
.

Supplementary data