Geochemical characteristics of dissolved heavy metals in Zhujiang River, Southwest China: spatial-temporal distribution, source, export flux estimation, and a water quality assessment

Study area

The Zhujiang River (Pearl River) basin is located in Southwest China (21°31′–26°49′N, 102°14′–115°53′E), which is the largest river flowing into the SCS with a cover area of 4.5 × 10⁵ km² and provides major water source for more than 30 million local population of southern China (Li et al., 2019; Zhen et al., 2016). The Zhujiang River flows through Yunnan, Guizhou, Guangxi, and Guangdong province with the elevation decreases from northwest to southeast (Fig. 1). Five principal reaches are defined along the main channel of Zhujiang River: the Nanpanjiang River (NPR), the Hongshuihe River (HSR), the Qianjiang River (QJR), Xunjiang River (XUR), and Xijiang River (XJR) (Fig. 1). Accordingly, the tributaries can also be divided into five parts. In addition, the significant tributary (Beipanjiang River, BPR) located in the upstream area, is discussed as an individual tributary here. The Zhujiang River basin is characterized by tropical to subtropical monsoon climate, with a mean annual temperature varying from 14 to 22 °C. The average annual precipitation ranges from 1,200 to 2,200 mm, where a majority of annual rainfall (∼80%) occurs during the wet season (April–September) (Han et al., 2018; Li et al., 2019). The Zhujiang River basin consists of various source rocks (Fig. 1). Permian and Triassic carbonate rocks with intercalated coal-bearing formations are extensively distributed in the mid-upper river reaches with an area account for 44% of the whole basin area. Jurassic detrital sedimentary rocks are distributed sporadically in the river source area and are intercalated fragmentarily in the middle and lower river reaches. The mid-lower reaches consist mostly of Precambrian metamorphic rocks and magmatic rocks (Han et al., 2018; Li et al., 2019, 2000). There are 24 large reservoirs/dams (with capacity exceed 10⁸ m³) and 212 medium reservoirs/dams (capacity from 10⁷ to10⁸ m³) located in the mid-lower reaches of the Zhujiang River (Han et al., 2018).

Sample collection and analysis

Two sampling campaigns from 81 sites represent varying hydrologic period settings throughout the Zhujiang River during July 2014 (high flow season) and January 2015 (low flow season) are conducted (Fig. 1). And 9, 3, 3, 5, and 5 sites are located in NPR, HSR, QJR, XUR, and XJR, while 5, 7, 7, 7, 11, and 19 sites are located in the corresponding tributaries of NPR, HSR, QJR, XUR, XJR, and BPR. More detailed information of the sampling points could be found in Table S1. Thus, a total of 162 river water samples are collected at a depth of approximate 15 cm. After collection, the water samples are immediately filtered in the field through 0.22 μm Millipore membrane filters. Water samples for heavy metals analysis are acidified to pH < 2 with ultra-purified HNO₃ and then sealed in a pre-cleaned polyethylene bottles and stored in a dark refrigerator before analysis. The pH, electrical conductivity (EC), dissolved oxygen (DO), and temperature (T) are immediately measured by a multi-parameter meter (MultiLine 3320; WTW, Weilheim, Upper Bavaria, Germany) in the field. Ten heavy metals, including V, Cr, Mn, Co, Ni, Cu, Mo, Cd, Ba, and Pb are analyzed by inductively coupled plasma-mass spectrometry (Elan DRC-e; Perkin Elmer, Waltham, MA, USA) at the Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences. Standard reference materials (GSB 04-1767-2004) are used to perform the method validation and quality control. All the samples and standards are analyzed in batches with a procedural blank. Recovery percentage ranges between 90.0% and 110.4%. Relative standard deviations for heavy metals are ∼±5%.

Statistical analysis and assessment methods

Kolmogorov–Smirnov test of data

The Kolmogorov–Smirnov test, as a non-parametric test, is generally used to data analysis when the sample size is small. Thus, the normal distribution test of our data is using Kolmogorov–Smirnov (K–S) statistics (Table 1). The test results show that the V, Cr, pH, EC, and DO are normally distributed during low flow season, and the V, Cr, Mo, pH, EC, DO, and temperature approach a normal distribution during high flow season. The K–S test results and large standard deviation values for the remaining dissolved heavy metals indicatives of that the average concentration might have been seriously impacted by the outliers, which related to water samples with tremendously high or low values. Thus, the median concentrations instead of arithmetic means for these heavy metals are used in our calculations. In addition, based on the K–S test results, the median concentrations of all studied dissolved heavy metals are used for comparisons, although the limited values in the guidelines are widely given as arithmetic means.

Distribution of heavy metals along the main stream

The distribution of each individual heavy metal along the main stream exhibits an extensive variation (Fig. 2). In the main stream water body, V concentration present a gradually increasing trend from upstream to downstream overall, with a higher concentration in the middle of the NPR reach. In contrast, Cr and Co exhibits a slowly decreasing trend from upstream to downstream, but perform a higher concentration in the middle of the NPR reach. Cu concentration is maintained at a low level, and present a very high value for only a few sampling points (10 times more than other points). Mo concentration gradually increase in the NPR reach and began to decline after entering the HSR reach, until the QJR reach become stable. There is a great fluctuation of Ba concentration in the upstream reach, with the highest concentration in the middle of NPR reach, and subsequently decrease to the lowest value in QJR reach, and then exhibit a slowly rising trend in the downstream reach (XUR and XJR). Pb concentration is fluctuated in a relatively wider range in NPR and HSR reach, and tended to stabilize in QJR to XJR reach. Mn, Ni, and Cd display obviously fluctuation along the main stream, with no significant pattern of distribution. Moreover, there are 24 large reservoirs/dams located in the mid-lower reaches along the main channel of the Zhujiang River (Han et al., 2018). The suspended particles, gravels that absorbed heavy metals would be separated from the water physically as water flow slows down in the reservoirs (Li et al., 2008), which could be considered as a potentially factor for the spatial variation of the dissolved heavy metals, and may result in the variation of concentration of some dissolved heavy metals in the mid-lower reaches.

Distribution of heavy metals in tributaries

The distribution of each individual heavy metals in tributaries of different river reaches is exhibited in Fig. 3. Compare the median concentrations of heavy metals in tributaries, the V concentration is lower in the upstream tributaries than the downstream tributaries (except T-NPR), while Cr and Co are higher in the upstream tributaries than the downstream tributaries. It is consistent with the V, Cr, and Co concentration distribution in the main stream. Inconsistent with the main stream, Ni concentration present a similar distribution pattern to Cr and Co in the tributaries of each reaches, which is higher in the upstream tributaries than the downstream tributaries. In addition, Mn concentration shows that the upstream tributaries are lower than the downstream tributaries, and the fluctuated range is obviously increased in the tributaries of the downstream region. The remaining heavy metals (Cu, Mo, Ba, Pb, and Cd) exhibit an extensive variation in the tributaries of each reaches, with no significant pattern, but it should be noted that the fluctuation of several heavy metals (Mo, Pb, and Cd) in the downstream tributaries are relatively more intense than in the upstream tributaries.

Concentrations of heavy metals in river water

The concentration of dissolved heavy metals are compared with the limited values for drinking water guidelines of China EPA (2006), World Health Organization (WHO) (2006), United States Environmental Protection Agency (EPA) (2003) (Table 1) and the surface water standard of China EPA (2002) (Table S2). All the concentrations (median) of heavy metals are within the limited values of three drinking water guidelines (except V, without limited values in any guidelines), and the Cr, Cu, Cd, and Pb concentrations (median) are of Grade I of surface water standard, which indicate that pretty clean water without heavy metals pollution in Zhujiang River. In addition, the maximum concentrations of Mn and Ni are lower than the limited values of WHO drinking water guideline but are approximately two times higher than the Chinese drinking water guideline, and the maximum concentrations of Mo in low flow season is higher than the Chinese drinking water guideline. These heavy metals with maximum concentrations exceed guideline values can be defined as pollutants, which could attribute to the relatively high anthropogenic input from local sampling sites. For instance, although the Cu concentrations at all sampling sites are well below the guideline values, but the Cu concentrations at the two sampling sites (site 1 and 2, close to the Zhanyi county) of NPR reach along the main stream are significantly higher than that of other sites (Fig. 2), which shows the significant impact of the anthropogenic input from the Zhanyi county (e.g., urban sewage discharge). Compare with the background values of the source area of Yangtze River (Table 1) (Zhang & Zhou, 1992), the concentrations (median) of V, Cr, Ni, Cu, and Cd in water within the Zhujiang River are significantly elevated, while Mn, Co, Mo, and Pb are lower than the source area of the Yangtze River. In low flow season, V, Cr, Ni, Cu, and Cd are 9.2, 26.4, 13.3, 1.4, and 2.9 times higher than the river water background of the Yangtze River. In high flow season, V, Cr, Ni, Cu, and Cd are 11.2, 28.9, 11.5, 1.2, and 2.2 times higher than the background values.

On a global scale, we compare our river samples with the data of other rivers both from China and other counties or regions worldwide (Gaillardet, Viers & Dupré, 2014) (Table S3). V, Cr, and Ni, concentrations of our study are higher than those worldwide rivers (except Seine River, a river severely affected by human activities in France). The Co and Mo concentrations are comparable to the worldwide rivers average concentration, but the Co concentration is much higher than that in Yellow River of China. Moreover, the Cu, Cd, Ba, and Pb concentrations are slightly lower than the world river average, while the Mn concentration is much lower than world river average.

Statistical analysis and source identification of heavy metals

Principal component analysis

Principal component analysis is carried out for heavy metal concentrations along the main channel of Zhujiang River to explore metal associations and their possible origins, three principal components (PC, eigenvalues > 1) are extracted in our study, including the eigenvalues, variance and communalities, are listed in Table 3. The PC 1 explain 22.59% of total variance and predominantly include V, Cr, Co, and Ba; the PC 2 explain 20.91% of total variance with significant loadings of Mn, Ni, and Cu; the PC 3 explain 18.23% of variance which is mainly contributed by Mo, Cd, and Pb. These three PCs totally account for 61.74% of the total variance, and are presented in a three-dimensional space, as shown in Fig. 4. Our results of overall PCs loadings (61.74%) are relatively lower than other studies, that is, PCs loadings for 14 heavy metals and 13 metals are 86.36 % and 79.31% (Meng et al., 2016; Wang et al., 2017), respectively. Previous studies have performed PCA for heavy metals in numerous of river system and attained different results (Li & Zhang, 2010; Wang et al., 2017; Xiao, Jin & Wang, 2014), we ascribe these variations to the different fluvial environments (i.e., watershed area, discharge, and land use) and different water variables.

Table 3:

Varimax rotated component matrix for dissolved heavy metals along the main channel of Zhujiang River (the significance of KMO and Bartlett’s sphericity test is <0.001).

Eigenvalues	2.26	2.09	1.82	Communalities
Variance (%)	22.59	20.91	18.23
Cumulative (%)	22.59	43.50	61.74
Variable	PC1	PC 2	PC 3
V	0.18	−0.22	−0.73	0.60
Cr	0.85	−0.02	−0.12	0.74
Mn	−0.20	0.81	−0.01	0.71
Co	0.84	0.04	0.12	0.72
Ni	0.10	0.67	0.03	0.46
Cu	−0.08	0.89	−0.04	0.80
Mo	0.32	−0.29	0.47	0.40
Cd	−0.03	−0.14	0.71	0.53
Ba	0.79	−0.15	−0.04	0.66
Pb	0.05	0.05	0.74	0.55

DOI: 10.7717/peerj.6578/table-3

Notes:

Extraction method: Principal component analysis.

Rotation method: Varimax with Kaiser normalization.

In general, the Cr, Co, Ba, and Mo are from natural sources of rock weathering and subsequent pedogenesis (Li & Zhang, 2010), and V is greatly impacted by anthropogenic activities such as mining and agricultural processes (Li et al., 2008). Considering the high concentrations of V and Cr (much higher than the background value of the Yangtze River, Table 1), and the most of elements in PC1 (V, Cr, Co, Ba) are lithophile elements (Krishna, Satyanarayanan & Govil, 2009), hence we attribute this component (PC 1) to the mixed sources of geologic and anthropogenic origins in the basin. The high concentrations of Cu and Ni in PC 2 are observed in Zhujiang River (compared to the background value), and Cu is used as markers of metal industries (Li & Zhang, 2010), while Ni is the common pollutant discharge from the electroplating industry and metal smelting. In combination with the negative loadings or weak positive loadings of Cu and Ni in PC 1 and the positive correlations between Cu and Ni (Table 2), the PC 2 with positive loadings on Cu (0.89) and Ni (0.67) can be attribute to anthropogenic origins in the basin. In addition, although Cd and Pb are major pollutants emitted from industrial wastes and automobile exhausts (Krishna, Satyanarayanan & Govil, 2009; Pekey, Karakaş & Bakogˇlu, 2004), but the concentrations of these two heavy metals and Mo are not high in the Zhujiang River, even lower than the background value of the Yangtze River. Therefore, Cd, Pb, and Mo in PC 3 may be greatly contributed by natural sources.

In spite of the results of PCA that suggest that some heavy metals (V, Cr, Ni, and Cu) are possibly caused by the anthropogenic inputs or the mixed sources of geologic and anthropogenic origins, the rest of the heavy metals with low concentrations below the background value are controlled by the natural sources or assuaged by the varying landscape setting and the relatively weak rock weathering processes. Compared to the polluted rivers such as Huaihe River (Cu (28.61 μg L⁻¹), Pb (97.83 μg L⁻¹), Ni (16.00 μg L⁻¹), Cr (19.7 μg L⁻¹)) (Wang et al., 2017), the concentration level of heavy metals in the Zhujiang River are fairly low (Table 1). Therefore, we hold that dilution has a great impact on the anthropogenic heavy metals (V, Cr, Ni, and Cu) within such a large catchment area. The dissolved heavy metals in Zhujiang River are mainly presented a natural sources signature.

Water quality and health risk in river water

The allowable limit of pH for drinking water ranges from 6.5 to 8.5 (Table 1). The pH values in 95.7% of the water samples are conformed to this regulation in our study. Based on PCA results, the weights of each heavy metal (w_i) are obtained and summarized in Table 4. Therefore, the calculated WQI values using Eq. (1) are shown in Table S1. In this study, the WQI values ranges from 1.5 to 43.9 and 1.3 to 27.9 for water samples during the low and high flow season, respectively. The calculated WQI values based on the median concentrations of heavy metals in the two seasons are 3.3 and 3.2. Only two sites (site 61 and 74) are evaluated as a relatively high WQI value (43.9 and 33.5) in low flow season. All the water samples can be defined as excellent water quality (the calculated WQI values are less than 50), indicative of that the natural water within the Zhujiang River drainage area is suitable for drinking, at least from the perspective of heavy metal pollution.

The HQ and HI values of the heavy metals via ingestion and dermal pathways for adults and children are calculated with the median concentrations based on Eqs. (2)–(6), respectively. As shown in Table 5, for both adults and children, the HQ_ingestion, HQ_dermal, and HIs for all heavy metals are smaller than 1, indicate that the metals investigated in the Zhujiang River basin are all below the hazard level (through oral ingestion and dermal absorption), and the health effects of these heavy metals are very limited. It should be noted that the HQ_ingestion and HQ_dermal values for children are relatively higher than adults, suggest that children are more risky under the heavy metals exposure. Moreover, at some sampling points with the maximal heavy metal concentrations, the HI of Mo (0.539 and 0.826 for adults and children, respectively; site 68) and Mn (0.573 for children; site 74) are relatively close to 1 in low flow season. Consequently, we conclude that Mo and Mn could be a potential non-carcinogenic risk to human health in some sites, especially in low flow season.

In spite of the risk, assessment indicates that widespread heavy metal pollutants are not in the Zhujiang River basin, which is similar to the results from WQI. Several studies have reported the adverse effects of heavy metals, such as the toxicity of Pb on the nervous and endocrine systems in the human body (Fang et al., 2014); the adverse effects of Cd in the renal system and bone injuries (Bertin & Averbeck, 2006); the effects of Mo on reproduction and foetal development (Vyskocil & Viau, 1999). Therefore, special measures should be taken to prevent the input of heavy metals into the Zhujiang River for the protection of its excellent water quality, and to provide a better water resources guarantees for socio-economic development within the whole basin.

Heavy metals export budget of Zhujiang River to the South China Sea

By taking the concentrations of the dissolved heavy metals in low flow season (October–March of the following year) and high flow season (April–September), and the discharge at the mouth of the Zhujiang River of the two seasons (River and Sediment Bulletin of China, http://www.mwr.gov.cn/sj/tjgb/zghlnsgb/), river fluxes of individual of heavy metal to the SCS are estimated that range from 8.6 (Pb) to 3,713.5 (Ba) tons during the hydrologic year of 2014–2015 (Table 6). It should be noted that 80%, 76%, 25%, 71%, 58%, 63%, 70%, 80%, 74%, and 63% of the V, Cr, Mn, Co, Ni, Cu, Mo, Cd, Ba, and Pb flux produced in the high flow season, which is mainly controlled by the high discharge (1.72 × 10¹¹ m³, account for 74.1% discharge of whole hydrologic year) in this season. Table 6 also presented the estimated heavy metal fluxes of Zhujiang River in 2002, which are calculated by the heavy metal concentrations from previous study (Ouyang et al., 2004) and the discharge in the corresponding year. Most of the estimated heavy metal fluxes are decreased from 2002 to 2015 for the Zhujiang River (except V) (Table 6). Thus, the environmental policy in last decade might reduce the anthropogenic heavy metal inputs to the Zhujiang River in China. Compared with literature values (Gaillardet, Viers & Dupré, 2014), the proportion of the annual estimated export fluxes of V, Cr, Mn, Co, Ni, Cu, Mo, Cd, Ba, and Pb of Zhujiang River in the global rivers are 2.7%, 6.0%, 0.01%, 0.4%, 1.7%, 0.3%, 0.7%, 0.4%, 0.4%, and 0.3%, respectively. Based on the annual river discharge of 3.74 × 10¹³ m³ year⁻¹ for global rivers (Xu & Han, 2009), the annual discharge of the Zhujiang River in the total annual discharge of the world’s rivers is estimated as 0.6%. The proportion of V, Cr, Ni, and Mo export fluxes of Zhujiang River in global rivers exceeds 0.6%, which suggests that the contribution of V, Cr, Ni, and Mo of Zhujiang River export to the marine system are higher than the world average level.

In this study, the concentrations of heavy metals in the estuary of the Zhujiang River present different degrees in the two seasons (Fig. 2). Given that most of the heavy metals export occurred in the rainy season, and there may be large uncertainty in estimation of heavy metal fluxes due to the limited temporal sampling interval in the Zhujiang River, particularly in the high flow season with a conspicuous variation of heavy metal concentrations after a storm event, our data from one sole sampling campaign may be incomplete. Therefore, high-frequency sampling (monthly, weekly even daily sampling) is necessary and is vital significance to precisely quantify heavy metal annual budget from the terrestrial rivers to the ocean.