Abstract

Spatial–temporal variations in 13 selected water quality parameters from four stations located in the stagnant Haihe River from 2012 to 2014 were analysed. Principal component analysis and cluster analysis were applied. The main latent anthropogenic factors affecting the water quality of Sanchakou, Sixin Bridge, Liulin, and Erdao Gate were combined sewer overflow, organic matter, domestic sewage, and agricultural diffuse source, respectively. External inputs mainly affected quality water in the summer–autumn season. By contrast, intrinsic biochemical processes were highly correlated with water quality in the winter–spring season. Ranges of total nitrogen (TN) and total phosphorus (TP) of four sampling sites measured 1.2 mg/L to 11.4 mg/L and 0.04 mg/L to 2.06 mg/L, respectively. TN/TP (mass ratio) was mainly between 9 and 23, indicating severely eutrophicated mainstream of the Haihe River and sufficient amounts of nutrients for phytoplankton growth and reproduction. Hence, dual nutrients control strategies should be implemented in this stagnant urban river.

INTRODUCTION

Water pollution has been one of the most significant threats to public health for a long time. Every year, 60,000 people in China die from diseases, such as liver and gastric cancers, which are caused by water pollution (Tao & Xin 2014). Previous studies have reported that up to 80% of urban rivers in China were contaminated to varying degrees, with several rivers appearing black, being malodorous and losing fish (Sheng et al. 2013). One quarter of 4,000 surveyed urban wastewater treatment plants failed to comply with quality controls, inciting public fear regarding health impacts (Tao & Xin 2014). Rapid industrialization, urbanization, and population growth mainly result in polluted urban rivers (Caraballo et al. 2011). Effective and efficient water management requires reliable information on water quality. However, practical urban water environments are complex and can be affected by a constant point sources that comprise industrial and municipal wastewater, surface runoff, precipitation, and pumped inflow. Therefore, systematic surveys and comprehensive assessment of water quality are essential in increasing knowledge on urban river water environment.

The Haihe River, which is located within the central political, economic and cultural area of China, serves not only as the main flood watercourse during rainy season but also as water supply source for drinking, industries and urban landscape recreation. However, the Haihe River features an extremely slow flow rate. With the development of agriculture and industry around the stagnant Haihe River, river quality gradually declines and serious algal blooms frequently result in ecological imbalance and deterioration of aesthetic landscape. A few studies have been reported on the sediment and water pollution of the Haihe River (Luo et al. 2011; Wu et al. 2011; Gao et al. 2012), and some ecological countermeasures have also been explored (Chi & Cai 2012; Chi & Yang 2012; Liu et al. 2014). However, only a few reports systematically analysed pollution sources in spite of their critical role in increasing knowledge on hydrology, geochemistry and pollution status of the Haihe River. Several parameters used to represent river water quality may become verbose and confusing (Ou et al. 2014). Converting multifaceted water quality data into simple information is necessary to ensure that knowledge is comprehensible and usable by the public (Al-Othman 2015).

Problems of data reduction and interpretation, characteristic changes in water quality parameters and indicator parameter identification can be addressed by using different methods, such as single factor index (SFI) (Al-Othman 2015), Nemerow's pollution index (NPI) (Adamu et al. 2015), artificial neutral networks (Singh et al. 2009), fuzzy logic (Yilmaz 2007), cluster analysis (CA) (Varol & Sen 2009) and principal component analysis (PCA) (Muangthong & Shrestha 2015). Among these methods, CA can indicate groupings of samples by linking intersample similarities and illustrating the overall similarity of variables in the data set. PCA can explain variance for large sets of intercorrelated variables and reduce variance into smaller sets of independent variables with minimum loss of original information. These methods provide invaluable insights to studies on complex surface water environment. This information is important to environmental management and protection.

In view of the aforementioned considerations, this study aimed to analyse spatial–temporal variations, classify surface water quality into different periods, identify major factors influencing water quality in selected periods by using statistical methods and conduct eutrophication assessment, which can lay the foundation for management and remediation of the Haihe River.

MATERIAL AND METHODS

Study area

The Haihe River basin is located in North China, measures 1,050 km long, features a total catchment area of 2,066 km2 and is politically important and culturally and economically advanced (Figure 1). The Haihe River serves as an important source of drinking, agricultural, industrial and aquaculture water. Similarly, this river presents an important key tourist attraction for over 10 million people. This study focuses on eastern basin of the Haihe River located at Tianjin City. The Haihe River measures 73 km, stems from Sanchakou and flows eastward to Bohai Bay and can be divided into three segments. The upstream part from Sanchakou to the outer loop spans 19 km and flows through the central area of Tianjin City. The middle stream measures approximately 18 km long and located in the countryside area between the outer loop and Erdao Gate. The downstream part lies within the Tanggu District. The upstream part comprises 131 storm water pumping stations, and 26.2 million m3·yr−1 of water is pumped into the Haihe River. Average annual atmospheric temperature measures approximately 12°C. Average temperature in January is the lowest at −4°C, whereas that in July is the highest at 26°C. Average annual evaporation capacity reaches 900–1,200 mm, and these during dry and wet seasons concentrate 54% and 46%, respectively. Annual precipitation amounts to 560–720 mm, and approximately 85% is concentrated during summer–autumn season. Approximately 30 million m3·yr−1 of water is transferred from the Luan River as supply source for ecological use.

Figure 1

Location map of the study area showing sampling sites.

Figure 1

Location map of the study area showing sampling sites.

Sampling and testing

Samples were typically collected monthly from Sanchakou to Erdao Gate during the years 2012 to 2014, as shown in Figure 1. Samples were collected according to the Environmental quality standards for surface water (GB18918-2002), and all measurements followed the national quality standards for surface water (APHA/AWWA/WEF 2012). Water temperature (T), pH and dissolved oxygen (DO) were determined in situ using the appropriate sensors. The other parameters were determined in a laboratory. The specific methods used were as follows: CODMn, acidic (alkaline) potassium permanganate method; Cr(VI), diphenylcarbohydrazide spectrophotometric method; fluoride (F), ion chromatography; total phosphorus (TP), ammonium molybdate spectrophotometric method; NH3-N, spectrophotometric method with salicylic acid.

Data treatment

Averaged data from March–May, June–August, September–November and December–February represent water quality for spring, summer, autumn and winter, respectively. PCA was conducted to convert a set of possible correlated variables into a set of linearly uncorrelated variables. Then, factor analysis (FA) was performed to reduce contribution of less significant variables and further simplify data structure that originated from PCA; such data can be achieved by rotating the axis (Muangthong & Shrestha 2015). New variables called varifactors (VFs) were constructed. Kaiser–Meyer–Olkin and Bartlett's sphericity tests were conducted to examine suitability of data for FA. CA was applied to reveal high internal homogeneity within clusters and high external heterogeneity between clusters. Here, the Ward's method was selected for sample classification. Ward's method uses analysis of variance approach to evaluate the distance between clusters in an attempt to minimise the error sum of squares (ESS). The ESS can be calculated as follows:  
formula
(1)
where xi refers to the score of ith individual (Ward 1963).

All multivariate analyses were conducted using the SPSS 22.0 statistical software (IBM, Armonk, NY, USA). Experimental data were standardised through z-scale transformation to avoid misclassification, which results from significant differences in data dimensionality.

RESULTS AND DISCUSSION

Spatial characterisation of the Haihe River mainstream

Variations in parameters along the Haihe River

Variations of parameters along the Haihe River were presented in the form of box plots (Figure 2). The line in the box represents the median, whereas bottom and top boxes showed the locations of the first and third quartiles. Whiskers correspond to lines extending from the top and bottom of the box, which represent the highest and lowest observations, respectively. CODMn, TP, NH3-N, alkalinity (ALK), Cl and F all exhibited an increasing trend along the flow direction, whereas DO demonstrated a decreasing trend along the flow direction. Average value of total nitrogen (TN) initially increased and then slightly decreased. TN and TP were the major contributors to eutrophication, which caused an excessive growth of algae that depleted DO through decomposition (Chang 2005). Decomposition of pollutants also consumed a considerable amount of DO. Anthropogenic pollution sources of F included runoff from agricultural land and urban areas, cement plants, fluorine chemical factories, phosphorus fertiliser plants, and smelters. F from Sanchakou to Erdao Gate increased gradually, and that of Erdao Gate measured higher than 1.5 mg/L. Therefore, industrial pollution and dispersive source increased along the river, with results indicating that anthropogenic disturbance exceeded the self-purification capability of the Haihe River.

Figure 2

The variation of parameters at four sampling sites along Haihe River. The maximum, minimum, median, upper and down quartile values are presented in the form of boxplot. Cr(VI): hexavalent chromium; F: fluoride; Cl: chloride; ALK: total alkalinity.

Figure 2

The variation of parameters at four sampling sites along Haihe River. The maximum, minimum, median, upper and down quartile values are presented in the form of boxplot. Cr(VI): hexavalent chromium; F: fluoride; Cl: chloride; ALK: total alkalinity.

The mainstream of the Haihe River was moderately saline with mean pH values of 8.27 for Sanchakou, 8.09 for Sixin Bridge and Liulin, and 7.98 for Erdao Gate; these values were within the recommended standard for drinking and domestic purposes (6.5–8.5). Previous studies reported that combined sewer overflow (CSO) outfalls are important point sources of pollution (Tao et al. 2014). As previously mentioned, Haihe River comprises 131 storm water pumping stations. During high-intensity storms, CSO is directly discharged from combined sewers or passes through a wastewater treatment facility to nearby water bodies. Large variations in NH3-N, NO2-N and TP were attributed to CSO during a rainfall periods. Meanwhile, large variation in water temperature was attributed to climate and seasonal influences on the study region.

FA based on sampling sites

In this study, FA was conducted on the normalised data, which were set separately for four different regions (Table 1). PCA of the four data sets yielded four principal components at each site with eigenvalues >1, which explained 77.41%, 77.77%, 69.47% and 77.08% of total variance. Significance of factor was characterised by the eigenvalues. High eigenvalues represent significant factors. Factor loadings were classified as weak, moderate, and strong, corresponding to values of 0.3–0.5, 0.5–0.75 and >0.75, respectively (Liu et al. 2003).

Table 1

Loadings of experimental variables on significant principal components for Sanchakou, Sixin Bridge, Liulin and Erdao Gate

Sanchakou
Sixin Bridge
Liulin
Erdao Gate
ParametersVF1VF2VF3VF4VF1VF2VF3VF4VF1VF2VF3VF4VF1VF2VF3VF4
ALK 0.93 0.14   0.95  0.23  0.89 0.17 0.33 −0.11 0.92 0.25   
Cl 0.91 0.17 −0.16  0.92   0.26 0.79 0.37 0.41  0.86 0.18 0.35  
0.85 0.29  0.18 0.87 −0.18 0.33 0.12 0.86 0.29 0.20  0.92  −0.15  
TN 0.74 −0.11   0.53 −0.45  −0.17 0.92 0.17 −0.17  0.38 0.84  −0.22 
CODMn 0.73 0.26  0.39 0.26 −0.15 0.15 0.86 0.21 0.44 0.49 0.41 0.74 −0.22 −0.40 −0.32 
0.68  0.53 0.14 −0.53 0.49 0.49 −0.15 −0.28 0.80  0.13 −0.19   0.89 
NO3-N  0.81 −0.24 −0.30 0.14 0.67 0.64 −0.19  0.11  0.93 0.11 −0.26 0.84  
Cr(VI) 0.14 0.78 −0.26  0.12  0.53   0.43  0.63 −0.15 0.38 −0.45 −0.47 
NH3-N 0.50 0.64  −0.22 0.54 −0.54 0.50 0.12 0.89  −0.14 0.37 0.33 0.81 −0.11 −0.30 
NO2-N   0.87  0.14  0.70 −0.40  0.83  −0.14 −0.25 −0.18 0.46 0.16 
DO 0.37 0.49 0.65 0.33  0.47 −0.39 0.67 0.12 0.14 0.85    0.74 0.36 
pH   −0.26 0.88 −0.15 0.87    −0.24 0.83 −0.12  −0.28 0.37 0.76 
TP 0.38  0.39 0.57 0.53 −0.43 0.62  0.72 −0.17  0.55 −0.11 0.78 −0.33  
Eigenvalue 5.38 1.83 1.68 1.18 5.22 2.34 1.33 1.22 5.19 2.19 1.66 1.47 4.44 2.80 1.61 1.17 
Proportion of variance/% 41.36 14.06 12.94 9.05 40.17 17.98 10.22 9.40 39.92 16.82 12.73 11.33 34.16 21.54 12.39 8.98 
Cumulative variance/% 41.36 55.42 68.36 77.41 40.17 58.14 68.37 77.77 39.92 56.74 69.47 80.81 34.16 55.71 68.10 77.08 
Sanchakou
Sixin Bridge
Liulin
Erdao Gate
ParametersVF1VF2VF3VF4VF1VF2VF3VF4VF1VF2VF3VF4VF1VF2VF3VF4
ALK 0.93 0.14   0.95  0.23  0.89 0.17 0.33 −0.11 0.92 0.25   
Cl 0.91 0.17 −0.16  0.92   0.26 0.79 0.37 0.41  0.86 0.18 0.35  
0.85 0.29  0.18 0.87 −0.18 0.33 0.12 0.86 0.29 0.20  0.92  −0.15  
TN 0.74 −0.11   0.53 −0.45  −0.17 0.92 0.17 −0.17  0.38 0.84  −0.22 
CODMn 0.73 0.26  0.39 0.26 −0.15 0.15 0.86 0.21 0.44 0.49 0.41 0.74 −0.22 −0.40 −0.32 
0.68  0.53 0.14 −0.53 0.49 0.49 −0.15 −0.28 0.80  0.13 −0.19   0.89 
NO3-N  0.81 −0.24 −0.30 0.14 0.67 0.64 −0.19  0.11  0.93 0.11 −0.26 0.84  
Cr(VI) 0.14 0.78 −0.26  0.12  0.53   0.43  0.63 −0.15 0.38 −0.45 −0.47 
NH3-N 0.50 0.64  −0.22 0.54 −0.54 0.50 0.12 0.89  −0.14 0.37 0.33 0.81 −0.11 −0.30 
NO2-N   0.87  0.14  0.70 −0.40  0.83  −0.14 −0.25 −0.18 0.46 0.16 
DO 0.37 0.49 0.65 0.33  0.47 −0.39 0.67 0.12 0.14 0.85    0.74 0.36 
pH   −0.26 0.88 −0.15 0.87    −0.24 0.83 −0.12  −0.28 0.37 0.76 
TP 0.38  0.39 0.57 0.53 −0.43 0.62  0.72 −0.17  0.55 −0.11 0.78 −0.33  
Eigenvalue 5.38 1.83 1.68 1.18 5.22 2.34 1.33 1.22 5.19 2.19 1.66 1.47 4.44 2.80 1.61 1.17 
Proportion of variance/% 41.36 14.06 12.94 9.05 40.17 17.98 10.22 9.40 39.92 16.82 12.73 11.33 34.16 21.54 12.39 8.98 
Cumulative variance/% 41.36 55.42 68.36 77.41 40.17 58.14 68.37 77.77 39.92 56.74 69.47 80.81 34.16 55.71 68.10 77.08 

Data lower than 0.1 were not presented.

For the data set pertaining to Sanchakou, VF1, which explained 41.36% of total variance, showed a strong positive loading on mineral-related substances (ALK, Cl, and F) and moderate positive loading on CODMn and TN. VF2, which explained 14.06% of the total variance, exhibited strong negative loading on nitrate (NO3) and Cr(VI). CSO may have played a key role after point source pollution was controlled effectively. VF3, which explained 12.94% of the total variance, presented strong positive loading on nitrite (NO2). VF4, which explained 9.05% of the total variance, showed a strong positive loading on pH. Latent factors that affected water quality of Sanchakou included mineral-related substances, CSO and physiochemical conditions (redox status and acid/alkaline conditions).

The data set on Sixin Bridge, VF1, which explained 40.17% of the total variance, indicated strong positive loading on mineral-related substances (ALK, Cl and F). VF2, which explained 17.98% of the total variance, resulted in strong positive loading on pH. VF3, which explained 10.22% of the total variance, showed moderate positive loading on NO2-N and TP and moderate negative loading on NO3-N. This factor also accounted for the anoxic status of the area. VF4, which explained 9.4% of the total variance, represented strong positive loading on CODMn. The latent factors that affected water quality of Sixin Bridge consisted of mineral-related substances, physiochemical conditions and organic matter.

For the data set related to Liulin, VF1, which explained 39.92% of total variance, indicated strong positive loading on nitrogen substances (TN and NH3-N) and mineral-related substances (ALK, Cl and F). VF2, which explained 16.82% of the total variance, corresponded to strong negative loading on NO2-N and T. Similar trends can be attributed to the inverse relationship of temperature and DO. High T easily saturated water with oxygen. Ammonia can only be converted into nitrite due to absence of oxygen. VF3, which explained 12.73% of total variance, depicted strong positive loading on DO and pH when consumption of ammonia and organic acids increased pH. Photosynthesis of algae can also increase DO and pH. VF4, which explained 11.33% of total variance, indicated strong negative loading on nitrate. Latent factors that affected water quality of Liulin constituted mineral-related substances, domestic sewage, redox status and photosynthesis.

For the data set regarding Erdao Gate, VF1, which explained 34.16% of total variance, showed strong positive loading on mineral-related substances (ALK, F and Cl) and moderate positive loading on CODMn. VF2, which explained 21.54% of total variance, exhibited strong positive loading on nutrient substances (TN, NH3-N and TP). Combination of VF1 and VF2 indicated that organic factors and nutrients were mainly contributed by nonpoint source pollution given that the Haihe River flows through a suburban agricultural area. VF3, which explained 12.39% of total variance, manifested strong positive loading on NO3 and moderate positive loading on DO. VF4, which explained 8.98% of total variance, showed strong positive loading on T and pH. This factor represents seasonal variation. Latent factors that affected water quality of Erdao Gate included mineral-related substances, diffuse source and seasonal variation.

Seasonal characterization in the mainstream of Haihe River

Variation of parameters in four seasons

The seasonal fluctuation of water quality is shown in Figure 3. Previous studies reported that seasonal variations of temperature, precipitation, and hydrological conditions strongly affected pollutant concentration in water (Xu et al. 2015). Pollutants in river systems typically originate from many transport pathways, which include storm water runoff, discharge from ditches and creeks, vadose zone leaching, groundwater seepage and atmospheric deposition, which are seasonal-dependent. Except for the slight decrease in TN from Liulin to Erdao Gate, all pollutant values increased from upstream to downstream in spring. Minimum average value of TP measured 0.26 mg/L in winter in Sanchakou, whereas its maximum average value totaled 0.68 mg/L in summer in Erdao Gate. Minimum average value of TN reached 2.83 mg/L in summer in Sanchakou, whereas its maximum average value amounted to 5.63 mg/L in winter in Liulin. Values of NO2-N and water temperature in summer–autumn season were higher than those in winter–spring season, whereas TN, NH3-N, DO, and mineral-related substances (F, Cl and ALK) exhibited the opposite trend. Although river quality was affected by CSO, water dilution in Luan River in summer–autumn season played an important role in decreasing concentration of pollutants. Results indicated complex biochemical variations in practical surficial river environment.

Figure 3

Seasonal distribution of parameters in the mainstream of Haihe River.

Figure 3

Seasonal distribution of parameters in the mainstream of Haihe River.

CA and FA in four seasons

CA enabled the grouping of river water samples based on similarities in chemical composition. Ward's method was employed for sample classification because of its small space-distorting effect. Ward's method also exhibits the smallest increase within sum of squares. Squared Euclidean distance was employed to measure dissimilarity.

The temporal dendrogram of the samples obtained by using Ward's method is shown in Figure 4. Two well-differentiated clusters can be observed. Cluster 1 mostly includes the summer–autumn samples from May to November, whereas Cluster 2 mostly includes the winter–spring samples from December to April. Clusters indicated highly heterogeneous quality of Haihe River in different seasons.

Figure 4

Dendrogram based on agglomerative hierarchical clustering for 23 sampling dates collected at Sanchakou, Sixin Bridge, Liulin and Erdao Gate from 2012 to 2014. ‘12 Jan.’ represented samples collected in January of 2012.

Figure 4

Dendrogram based on agglomerative hierarchical clustering for 23 sampling dates collected at Sanchakou, Sixin Bridge, Liulin and Erdao Gate from 2012 to 2014. ‘12 Jan.’ represented samples collected in January of 2012.

FA was separately conducted on Clusters 1 and 2 to elucidate the intrinsic relationships between parameters (Table 2). Four VFs were selected in Cluster 1, with a cumulative proportion of 76.71%. VF1 explained 37.91% of variance and was highly contributed by mineral-related substances (F, Cl and ALK). VF2 explained 17.1% of variance and included nutrient substances (NH3-N, TN and TP). VF3 explained 13.11% of variance and was positively contributed by Cr(VI) and T and negatively contributed by NO3-N. VF4 explained 8.59% of total variance of original data and was highly contributed by pH and DO. In Cluster 2, three VFs were selected, with a cumulative proportion of 64.64%. VF1 explained 36.77% of total variance and was positively contributed by NH3-N and TP and negatively contributed by NO3-N, pH and DO. VF2 explained 15.46% of variance and was highly contributed by mineral-related substances (Cl, ALK and F). VF3 explained 12.41% of variance and was positively contributed by T.

Table 2

FA of the Haihe River based on CA

Summer–autumn (cluster 1)
Winter–spring (cluster 2)
ParametersVF1VF2VF3VF4VF1VF2VF3
F 0.91 0.27 0.13 – 0.44 0.86 – 
Cl 0.85 0.40 – – – 0.91 – 
ALK 0.81 0.50 – 0.19 0.19 0.89 0.21 
NO2-N 0.69 – – −0.35 – – 0.47 
NH3-N 0.25 0.90 0.20 – 0.77 0.35 0.30 
TN 0.21 0.80 −0.34 – 0.48 0.11 0.60 
TP 0.38 0.68 0.44 – 0.69 0.11 0.34 
NO3-N – −0.17 0.89 0.15 0.83 – 0.20 
Cr(VI) 0.16 −0.13 0.73 −0.17 0.60 0.16 – 
– – 0.65 0.16 −0.19 0.23 0.77 
COD 0.34 0.47 0.56 – – 0.61 −0.55 
pH – 0.10 – 0.92 0.73 – – 
DO – −0.17 – 0.82 0.64 −0.25 −0.15 
Eigenvalue 4.92 2.22 1.70 1.11 4.78 2.01 1.61 
Proportion of variance/% 37.91 17.10 13.11 8.59 36.77 15.46 12.41 
Cumulative proportion/% 37.91 55.01 68.12 76.71 36.77 52.23 64.64 
Summer–autumn (cluster 1)
Winter–spring (cluster 2)
ParametersVF1VF2VF3VF4VF1VF2VF3
F 0.91 0.27 0.13 – 0.44 0.86 – 
Cl 0.85 0.40 – – – 0.91 – 
ALK 0.81 0.50 – 0.19 0.19 0.89 0.21 
NO2-N 0.69 – – −0.35 – – 0.47 
NH3-N 0.25 0.90 0.20 – 0.77 0.35 0.30 
TN 0.21 0.80 −0.34 – 0.48 0.11 0.60 
TP 0.38 0.68 0.44 – 0.69 0.11 0.34 
NO3-N – −0.17 0.89 0.15 0.83 – 0.20 
Cr(VI) 0.16 −0.13 0.73 −0.17 0.60 0.16 – 
– – 0.65 0.16 −0.19 0.23 0.77 
COD 0.34 0.47 0.56 – – 0.61 −0.55 
pH – 0.10 – 0.92 0.73 – – 
DO – −0.17 – 0.82 0.64 −0.25 −0.15 
Eigenvalue 4.92 2.22 1.70 1.11 4.78 2.01 1.61 
Proportion of variance/% 37.91 17.10 13.11 8.59 36.77 15.46 12.41 
Cumulative proportion/% 37.91 55.01 68.12 76.71 36.77 52.23 64.64 

In the summer–autumn season, four VFs represented mineral-related substances, eutrophication degree, CSO and photosynthesis. Although Cr(VI) level in this study was within standards, its concentration increased from Sanchakou to Erdao Gate (Figure 2). Cr(VI) applied to the Haihe River and precipitated during storm events. Then, Cr(VI) was released from sediments under high water temperature on sunny days. Three VFs represented eutrophication degree, mineral-related substances and water temperature in winter–spring season. VF1 can be explained by high levels of intrinsic nutrients that consumed substantial amounts of oxygen and underwent anaerobic fermentation, which increased ammonia and TP. Water pH values decreased due to hydrolysis of these pollutants. Eutrophication posed a severe problem all year round. However, external inputs mainly affected water quality in the summer–autumn season, whereas intrinsic biochemical processes were highly correlated with water quality in the winter–spring season.

Analysis of nitrogen and phosphorus

Nitrogen and phosphorus represent the two most important basic elements that limit algal growth (Xu et al. 2011). As shown in Figure 5, average TN/TP ratio (w/w) decreased from Sanchakou to Erdao Gate. By contrast, average TN/TP ratio (w/w) decreased from spring to summer and then increased gradually in subsequent seasons. TN/TP ratio revealed the limiting nutrient for phytoplankton growth in water bodies. Nitrogen-deficient growth occurs when TN/TP < 9 (mass), whereas phosphorus limitation occurs when TN/TP > 16, and phosphorus-deficient growth occurs when TN/TP > 23 (mass) (Xu et al. 2015). In this study, most of the values ranged from 9 to 23, indicating that neither nitrogen nor phosphorus was the limiting factor for phytoplankton growth.

Figure 5

Mass ratio of TN to TP.

Figure 5

Mass ratio of TN to TP.

The results of the analysis of the proportion of bioavailable inorganic nitrogen and the component of TN are shown in Figure 6. Over 63% of nitrogen was inorganic, whereas the proportion of organic nitrogen was relatively small. Percentage of ammonia also increased from upstream to downstream. The proportion of NO3 was larger than those of ammonia in Sanchakou, whereas the opposite phenomenon occurred in Erdao Gate. In general, ammonia is subjected to nitrification in natural river because of water self-purification. However, the percentage of ammonia increased, whereas that of NO3 decreased from Sanchakou to Erdao Gate. Opposing trends of ammonia and NO3 indicated that anthropogenic disturbance was exacerbated along the watercourse and exceeded self-purification capability of the Haihe River.

Figure 6

Component of TN at four sampling sites.

Figure 6

Component of TN at four sampling sites.

Overall, the Haihe River is prone to eutrophication. Considering the saline characteristic of Haihe River mainstream, saline-resistant plants should be selected as pioneer species during ecological remediation of the river (Zhao et al. 2014). An algal density investigation should also be included in a systematic monitoring scenario. Additional studies should be conducted to recognise the relationship between eutrophication and algal bloom for implementation of cost-effective, environment-friendly and comprehensive ecological remediation measures.

CONCLUSION

Nitrogen and phosphorus are two significant factors that lead to eutrophication and thus should be controlled simultaneously. Latent anthropogenic factors that affect water quality of Sanchakou, Sixin Bridge, Liulin, and Erdao Gate comprise CSO, organic matter, domestic sewage and agricultural diffuse source, respectively. External inputs mainly affect water quality during summer–autumn season. By contrast, water quality is highly correlated with intrinsic biochemical processes in the winter–spring season. Further studies should be conducted on N/P ratio of the Haihe River.

ACKNOWLEDGEMENTS

This work was financially supported by Major Science and Technology Project of Water Pollution Control and Management in China (No. 2014ZX07203-009).

REFERENCES

REFERENCES
Adamu
,
C. I.
,
Nganje
,
T. N.
&
Edet
,
A.
2015
Heavy metal contamination and health risk assessment associated with abandoned barite mines in Cross River State, southeastern Nigeria
.
Environmental Nanotechnology Monitoring & Management
3
,
10
21
.
Al-Othman
,
A. A.
2015
Evaluation of the suitability of surface water from Riyadh mainstream Saudi Arabia for a variety of uses
.
Arabian Journal of Chemistry
(
18
),
1
7
.
APHA/AWWA/WEF
2012
Standard Methods for the Examination of Water and Wastewater
, 22nd edn.
American Public Health Association/American Water Works Association/Water Environment Federation
,
Washington, DC
,
USA
.
Caraballo
,
M. A.
,
Macías
,
F.
,
Rötting
,
T. S.
,
Nieto
,
J. M.
&
Ayora
,
C.
2011
Long term remediation of highly polluted acid mine drainage: a sustainable approach to restore the environmental quality of the Odiel river basin
.
Environmental Pollution
159
(
12
),
3613
3619
.
CSEPA
2002
Environmental Quality Standards for Surface Water (GB18918-2002)
, 3rd edn.
China State Environmental Protection Administration
,
Beijing
,
China
.
Gao
,
L.
,
Shi
,
Y.
,
Li
,
W.
,
Liu
,
J.
&
Cai
,
Y.
2012
Occurrence, distribution and bioaccumulation of antibiotics in the Haihe River in China
.
Journal of Environmental Monitoring
14
(
4
),
1247
1254
.
Liu
,
C. W.
,
Lin
,
K. H.
&
Kuo
,
Y. M.
2003
Application of factor analysis in the assessment of groundwater quality in a blackfoot disease area in Taiwan
.
Science of the Total Environment
313
(
1–3
),
77
89
.
Luo
,
Y.
,
Xu
,
L.
,
Rysz
,
M.
,
Wang
,
Y.
,
Zhang
,
H.
&
Alvarez
,
P. J. J.
2011
Occurrence and transport of tetracycline, sulfonamide, quinolone, and macrolide antibiotics in the Haihe River Basin, China
.
Environmental Science & Technology
45
(
5
),
1827
1833
.
Ou
,
H. S.
,
Wei
,
C. H.
,
Deng
,
Y.
,
Gao
,
N. Y.
,
Ren
,
Y.
&
Hu
,
Y.
2014
Principal component analysis to assess the efficiency and mechanism for enhanced coagulation of natural algae-laden water using a novel dual coagulant system
.
Environmental Science and Pollution Research
21
(
3
),
2122
2131
.
Sheng
,
Y.
,
Qu
,
Y.
,
Ding
,
C.
,
Sun
,
Q.
&
Mortimer
,
R. J. G.
2013
A combined application of different engineering and biological techniques to remediate a heavily polluted river
.
Ecological Engineering
57
(
57
),
1
7
.
Singh
,
K. P.
,
Basant
,
A.
,
Malik
,
A.
&
Jain
,
G.
2009
Artificial neural network modeling of the river water quality-a case study
.
Ecological Modelling
220
(
6
),
888
895
.
Tao
,
T.
&
Xin
,
K.
2014
A sustainable plan for China's drinking water
.
Nature
511
(
7511
),
527
528
.
Ward
,
J. H.
1963
Hierarchical grouping to optimize an objective function
.
Journal of American Statistical Association
58
(
301
),
236
244
.
Wu
,
M.
,
Huang
,
S.
,
Wen
,
W.
,
Sun
,
X.
,
Tang
,
X.
&
Scholz
,
M.
2011
Nutrient distribution within and release from the contaminated sediment of Haihe River
.
Journal of Environmental Sciences
23
(
7
),
1086
1094
.
Xu
,
D.
,
Gao
,
Z.
,
Zhang
,
X.
,
Qi
,
Z.
,
Meng
,
C.
,
Zhuang
,
Z.
&
Ye
,
N.
2011
Evaluation of the potential role of the macroalga Laminaria japonica for alleviating coastal eutrophication
.
Bioresource Technology
102
(
21
),
9912
9918
.
Xu
,
Y.
,
Xie
,
R. Q.
,
Wang
,
Y. Q.
&
Sha
,
J.
2015
Spatio-temporal variations of water quality in Yuqiao Reservoir Basin, North China
.
Frontiers of Environmental Science & Engineering
9
(
4
),
649
664
.
Yilmaz
,
I.
2007
Fuzzy evaluation of water quality classification
.
Ecological Indicators
7
(
3
),
710
718
.