Abstract

Tail water from wastewater treatment plants (WWTP) serves as a major supplementary water source for scenic water bodies, whose water quality is one of the major focuses of public and scientific inquiries. This study investigated the temporal and spatial variations in water quality of Tangxihe River, a eutrophic urban river receiving tail water from a nearby WWTP in Hefei City, using the single-factor index (SFI) and principal component analysis (PCA). The results of SFI indicated that the most important parameters responsible for low water quality were total nitrogen (TN) and ammonia (NH4+-N). PCA showed that tail water from the WWTP greatly reduced water quality, as demonstrated by the significantly increased SFIs and integrated principal component values (F values) of the sampling points around the drain outlet of the WWTP (T3, T4 and T5). The sampling points located at the upstream of the river (T1) and up the water-gate of Chaohu Lake (T6) had negative F values, indicating relatively higher water quality. In addition, the season had a significant effect on the water quality of the river. Moreover, we discuss measures to improve the water quality of urban rivers in order to maintain their ecological functions.

INTRODUCTION

Urban rivers play a significant ecological, economic, and social role in the sustainability of human society. However, rapid urbanization and the consequent increase of water demand and wastewater discharge have caused severe water shortages and pollution in nearby rivers. In a number of water-scarce cities in China, effluent from wastewater treatment plants (WWTP) has become a primary supplementary water source for scenic rivers (Lyu et al. 2016). Unfortunately, given the relatively low discharge standard of pollutants for WWTP (Table S1, see Supplementary Material, available with the online version of this paper), the effluent likely becomes a pollution source for these rivers. Thus, changes in the water quality of these rivers increasingly cause great concern among the public and scientific researchers.

The availability of suitable tools to evaluate water quality is essential for successful risk assessment and management of urban rivers, and considerable effort has been made in the development of comprehensive water quality assessment methods (Wang et al. 2017a). Several methods have been developed to assess water quality, including the methods of single-factor index (SFI), biological indices (Pham 2017), comprehensive pollution index (Effendi 2016), the Nemerow pollution index (Chen et al. 2016), principal component analysis (PCA) (Ali & Khairy 2016), and comprehensive water quality identification index (Wang et al. 2017b). The SFI method, based on the Environmental Quality Standards for Surface Water (GB3838-2002) (EPA of China 2002), can be used to directly understand the relationship between water quality status and the assessment standard and identify the main factors leading to water pollution. The PCA method, which interprets the whole data set, provides data reduction and summarizes the statistical correlation among water quality constituents with minimum loss of the original information (Ouyang et al. 2006; Zhang et al. 2014).

Using the SFI and PCA methods, the objectives of this study are to assess the seasonal variation of the water quality of an urban river that receives the discharge of a WWTP, identify the main indicators responsible for the pollution condition of the river and reveal the main polluted area.

MATERIALS AND METHODS

Study site and sampling points

Tangxihe River, as a tributary of Chaohu Lake and an important landscape river in Hefei City, has a basin area of 50 km2. It originates from Nanyan Lake, with a depth of less than 3 metres. To investigate the temporal variation in water quality of the river, we selected six points (T1, T2, T3, T4, T5 and T6) along the river as water sampling sites (Figure 1). Due to insufficient headwater flow, the effluent of a nearby domestic WWTP (Tangxihe WWTP) is discharged into Tangxihe River as a major supplementary water source at the site between T3 and T4 at a flow rate of 30,000 m3/d. The water quality parameters of the effluent from spring of 2015 to summer of 2016 are shown in Table S1 (see in Supplementary Material, available with the online version of this paper).

Figure 1

Sampling sites of Tangxihe River.

Figure 1

Sampling sites of Tangxihe River.

Site T1 is located at the upstream of Tangxihe River; Site T2 is close to a residential area receiving nearby storm runoff; Site T3 and Site T4 are close to the point where the effluent from Tangxihe wastewater treatment plant is discharged into the river; Site T5 is near Baohe Road and receives some road runoff; and Site T6 is located at the downstream and before the water-gate of Chaohu Lake.

Water sampling and analysis

The water samples were collected from the six sites every month from March 2014 to February 2016. The physicochemical parameters, including pH, dissolved oxygen (DO), total suspended solids (TSS), and turbidity were monitored in situ with a portable water quality analyser (Professional Plus, YSI, USA), portable turbidimeter (2100Q, HACH, USA) and TSS portable instrument (HACH, CO, USA). Water samples were taken to the laboratory and analysed as soon as possible thereafter. The total nitrogen (TN), ammonia (NH4+-N), nitrite (NO2-N), nitrate (NO3-N), total phosphorus (TP), inorganic phosphorus (PO43−-P) and chemical oxygen demand (COD) were determined using the methods proposed in Standard Methods for Water and Wastewater Monitoring and Analysis.

Water quality assessment methods

Single-factor index method

To evaluate the water quality of Tangxihe River with the integration of a regulatory standard, the SFI method was used to compare the selected parameters (COD, TN, TP, NH4+-N and DO) with the threshold concentrations in Environmental Quality Standard for Surface Water (GB3838-2002) (EPA of China 2002). The threshold values of five grades of the standard are listed in Table S2 (see in Supplementary Material, available with the online version of this paper). The water quality standard for Tangxihe River is Grade V, so we selected Grade V as the target standard with which the determined parameters were compared. The SFI was calculated by Equation (1): 
formula
(1)
where Pi represents the SFI value of water quality of indicator i, Ci (mg/L) represents the measured value of indicator i, and Cti (mg/L) represents the threshold value of Grade V for indicator i under Environmental Quality Standard for Surface Water. Since DO values generally vary in an opposite direction to the other indicators (higher DO values indicate higher water quality), DO concentrations need to be transformed into their inverses before the SFI calculation.

Principal component analysis method

To combine the multiple indicators into an independent comprehensive index to represent the original data, an assessment method based on PCA was used. First, due to their different dimensions, all original indicators need to be standardized using the following formula: 
formula
(2)
where Cs is the standardized value of the indicator, C0 (mg/L) is the original value of the indicator, Cm (mg/L) is the mean value of the original indicators, and STD is the standard deviation of all original values of the indicators.
Then, a Pearson correlations matrix was prepared within the standardized indicators using SPSS 17.0 (SPSS Inc., Chicago, USA). According to the cumulative contribution of variance and the eigenvalue of the correlations matrix, the principal components (P1, P2, P3Pi) can be extracted, with a cumulative contribution of variance more than 65% and eigenvalue more than 1. The coefficient between the principal component Pi and each standardized indicator (Cof) was obtained from the component score coefficient matrix, and then the F value of the principal component Pi was calculated by Equation (3): 
formula
(3)
where Fi is the F value of the principal component Pi, n is the number of standardized indicators, Cj is the value of standardized indicator j, and Cofij is the coefficient between the principal component Pi and standardized indicator j.
Finally, the F is calculated as follows: 
formula
(4)
where F is the integrated F value of all standardized indicators, λi is the eigenvalue percentage of principal component i of the total eigenvalues of all extracted principal components.

Statistical analysis

The average SFI and F values of each season were calculated. Two-way analysis of variance (ANOVA) was used to test the differences in SFIs and F values between different sampling sites and seasons using SPSS 17.0 (SPSS Inc., Chicago, USA). The post hoc tests that compare the differences among sampling sites and seasons were conducted using the Student–Newman–Keuls method. Homogeneity of variances was ensured with a Levene's test before the ANOVA tests were performed. All significances are reported at the P < 0.05 level, unless otherwise stated.

RESULTS AND DISCUSSION

Water quality assessment using SFI method

SFI represents the relative pollution status compared with the target water quality standard. The SFI values of COD, TN, TP, NH4+-N and DO of the six sampling sites in different seasons are shown in Figure 2.

Figure 2

Single factor indices (SFI) of (a) total nitrogen (TN), (b) total phosphorus (TP), (c) ammonia (NH4+-N), (d) chemical oxygen demand (COD), and (e) dissolved oxygen (DO) of the six sampling sites in different seasons.

Figure 2

Single factor indices (SFI) of (a) total nitrogen (TN), (b) total phosphorus (TP), (c) ammonia (NH4+-N), (d) chemical oxygen demand (COD), and (e) dissolved oxygen (DO) of the six sampling sites in different seasons.

Except for DO, the SFI of all indicators showed significant differences among the sampling sites (P < 0.01, Table 1), and T3, T4 and T5 (4.12, 4.52 and 3.99, respectively) generally had higher SFIs than T1, T2 and T6 (2.22, 3.07 and 2.43, respectively). For TP and NH4+-N, the SFIs followed a similar trend to TN. However, the post hoc test of COD among the sampling sites demonstrated the mean SFI of COD at T6 was comparable to that at T3, T4 and T5. Given that the effluent from the WWTP joins the river at the location between T3 and T4, we can interpret that the tail water considerably impacted the river's water quality.

Table 1

Result of two-way ANOVA for the single factor index (SFI) of total nitrogen (TN), total phosphorus (TP), chemical oxygen demand (COD), ammonia (NH4+-N), dissolved oxygen (DO) and F value in Tangxihe River with the sampling sites and seasons as the two factors

IndicatorSite
Season
Site × Season
FSig.FSig.FSig.
SFI of TN 6.65 0.00 18.323 0.00 0.408 0.975 
SFI of TP 6.484 0.00 5.038 0.002 0.602 0.871 
SFI of NH4+-N 12.64 0.00 7.184 0.00 1.12 0.342 
SFI of COD 4.734 0.00 1.753 0.158 0.61 0.864 
SFI of DO 1.78 0.12 0.644 0.588 0.878 0.59 
Factor (F) 11.673 0.00 19.744 0.00 0.709 0.773 
IndicatorSite
Season
Site × Season
FSig.FSig.FSig.
SFI of TN 6.65 0.00 18.323 0.00 0.408 0.975 
SFI of TP 6.484 0.00 5.038 0.002 0.602 0.871 
SFI of NH4+-N 12.64 0.00 7.184 0.00 1.12 0.342 
SFI of COD 4.734 0.00 1.753 0.158 0.61 0.864 
SFI of DO 1.78 0.12 0.644 0.588 0.878 0.59 
Factor (F) 11.673 0.00 19.744 0.00 0.709 0.773 

Sig. represents the significance level.

The effect of season on the water quality of the river was demonstrated by the results of two-way ANOVA (Table 1). The SFI of TN significantly differed among seasons (P < 0.01), with the lowest mean value of 1.95 ± 1.28 in fall, and highest mean value of 4.03 ± 1.43 and 4.83 ± 3.27 in spring and winter, respectively. With respect to TP, season also had a notable effect on its SFIs, showing the lowest value of 0.46 ± 0.44 in fall and similar levels in other seasons. The SFI of NH4+-N reached its highest level in winter, with the value of 2.39 ± 1.41. For other indicators, including COD and DO, the differences in their SFIs among different seasons were not significant (P > 0.05).

The SFI values, which represent the gap between the water quality standard thresholds and the real water quality, had distinct ranges for different indicators. The SFI of TN and NH4+-N, with the mean values of 3.39 ± 2.41 and 1.76 ± 1.45, were notably higher than that of TP (0.86 ± 0.65), COD (0.65 ± 0.40) and DO (0.65 ± 0.54). The significant difference among indicators implies that TN and NH4+-N were the main factors responsible for the low real water quality compared with the target water quality.

Nitrogen and phosphorus pollutants primarily come from industrial wastewater (untreated or without thorough treatment), domestic wastewater, organic waste, livestock or poultry droppings, agricultural fertilizer, etc. At the location of sampling points T3 and T4 (close to a sewage treatment plant discharging industrial wastewater) and T5 (near the Baohe Road, which receives road runoff), the concentration of NH4+-N and P showed a dramatic elevated trend.

Furthermore, the higher SFI values of all indicators indicate that the pollution condition of the river was worse in spring and winter than in fall. Apart from the seasonal pattern of tail water from the WWTP, the possible factors driving this seasonal variation include the following. (1) The increased algae biomass in fall. For TN, for example, Wang et al. (2011) found that in the growing process of algae, the nutrient concentration gradient between the pore water and overlying water increased, with the absorption of a large amount of nutrients. This promoted the release of nutrients from the pore water to overlying water and ultimately led to the increase in concentration of TN during the winter period. (2) Hydrological force and water temperature. The river's inflow water quantity in the flood and non-flood seasons, along with its temperature, are both important factors influencing water quality (Rothenberger et al. 2014). In the summer and fall, the increased rainfall and temperature not only dilutes the pollutants but also promotes their degradation by microorganisms. (3) The composition of both phytoplankton and zooplankton. Ye et al. (2006) showed that the NH4+-N concentration exceeded the target threshold significantly in the summer, possibly due to the release of NH4+-N from sediment and the role played by microbes. In addition, it showed that one possible cause for high TP levels was the release of phosphorus absorbed by Fe from sediment into the overlying water during the summer (Ingall & Jahnke 1994).

Water quality assessment using PCA method

The value of F reflects the comprehensive pollution state of different sampling points, with a higher F value signifying lower water quality. The values of F at different sampling points in different seasons are shown in Figure 3.

Figure 3

The integrated value of principal component analysis (F) of different sampling points in different seasons.

Figure 3

The integrated value of principal component analysis (F) of different sampling points in different seasons.

Since a higher F value represents lower water quality, and based on the classification of pollution degree, the water quality of the sampling points T3, T4 and T5 always showed a serious pollution situation in spring, summer and winter, while the conditions of T1 and T6 were less severe. A possible reason for this observed variation is that the WWTP effluent enters the river near the location between T3 and T4, eventually leading to the serious pollution situation, which is consistent with the result of the SFI analysis. Furthermore, the location of T1, has good water quality readings. Since T2 was closer to a residential area, and T5 receives drainage from the road, these points could easily be influenced by human activities and storm pollution, which are considered to be the main sources of NH4+-N (Ikem et al. 2013). With respect to T6, one possible reason for its changes in water quality over time may be the occurrence of an algae bloom. Overall, the specific results of the sampling points T3, T4 and T5 may be related to their location near the WWTP, results that are supported by Olsen et al. (2012).

In addition, the F values were negative in fall while keeping the highest value in winter of all sampling points (seen in Figure 3), indicating the water quality of all sampling points was in good status in fall while seriously polluted in winter. It means the F value changes showed a significant seasonal pattern (similar to SFIs, seen in Table 1, P < 0.01), which is consistent with the study by Sharma & Chhipa (2016). The mean F of each season corresponded to the following order: fall (−0.27 ± 0.34) < spring (−0.029 ± 0.46) = summer (−0.02 ± 0.50) < winter (0.35 ± 0.55). The observed high water quality in fall and low water quality in winter may be due to rainfall (average 65.00 mm in fall and 39.33 mm in winter) and water temperature (average 22°C in fall and 8.33°C in winter), which is supported by Helena et al. (2000).

Based on the seasonal and spatial variations of water quality, the following measures to improve the water quality of Tangxihe River are proposed (including considerations of seasonality): (1) controlling the storm event runoff, which is the main source of NH4+-N and P; (2) improving the effluent standard for WWTPs, given the considerable pressure tail water exerts on river water quality; (3) implementing in situ approaches to restore vegetation and support the aquatic ecosystem, which is favourable for maintaining water quality (Wu et al. 2013). Regarding the species of vegetation to be restored, various studies have been conducted that stress the role of emergent and submerged plants, e.g., Acorus calamus, Iris tectorum, Canna generalis and C. generalis, in removing nitrogen and phosphorus pollutants and controlling algal bloom (Li 2011).

CONCLUSIONS

Water quality of the urban river receiving tail water as water source could be considerably impacted by the tail water quality. The main factors responsible for the pollution condition of Tangxihe River were TN and NH4+-N. The water quality was always the lowest at the sampling points close to the effluent outlet of the WWTP (T3, T4 and T5), while those located at the upstream of the river (T1) and before the water-gate of Chaohu Lake (T6) were higher. Hence, to improve the water quality of the river, the supplementary water resource (water from WWWTP) should be controlled in its quality. Moreover, Tangxihe River was most polluted in winter and least polluted in fall. Therefore, the measure of re-vegetating the surrounding area with hardy plants of strong repair ability should be taken to improve the Tangxihe River water quality.

ACKNOWLEDGEMENTS

This study was supported by the projects of National Major Programme of Science and Technology (2011ZX07303-001). The authors are also grateful to Mr Xiaohang Ma, Mr Shengyang Rui, Ms Jing Zhao, Ms Linya Shen, Ms Jia Xie and Ms Shan Huang for water monitoring.

REFERENCES

REFERENCES
Chen
R. H.
,
Li
F. P.
,
Zhang
H. P.
,
Jiang
Y.
,
Mao
L. C.
,
Wu
L. L.
&
Chen
L.
2016
Comparative analysis of water quality and toxicity assessment methods for urban highway runoff
.
Science of the Total Environment
553
,
519
523
.
Environmental Protection Agency of China
2002
Environmental Quality Standards for Surface Water (GB 3838–2002)
.
Helena
B.
,
Pardo
R.
,
Vega
M.
,
Barrado
E.
,
Fernandez
J. M.
&
Fernandez
L.
2000
Temporal evolution of groundwater composition in an alluvial aquifer (Pisuerga River, Spain) by principal component analysis
.
Water Research
34
(
3
),
807
816
.
Ikem
A.
,
Broz
B.
,
Garth
J.
,
Tesfaye
S.
&
Chung-Ho
L.
2013
Evaluation of surface water quality impacted by sewage overflows from animal and residential lagoon systems using principal component analysis
.
Environmental & Analytical Toxicology
3
(
7
),
1000197
.
Li
L.
2011
Growth Characteristics and Roles of 11 Species of Wetland Plants in Polluted Water Purification
.
Master's thesis
,
Jinan University
,
China
.
Lyu
S. D.
,
Chen
W. P.
,
Zhang
W. L.
,
Fan
Y. P.
&
Jiao
W. T.
2016
Wastewater reclamation and reuse in China: opportunities and challenges
.
Journal of Environmental Sciences
39
,
86
96
.
Ouyang
Y.
,
Nkedi-Kizza
P.
,
Wu
Q. T.
,
Shinde
D.
&
Huang
C. H.
2006
Assessment of seasonal variations in surface water quality
.
Water Research
40
(
20
),
3800
3810
.
Pham
L.
2017
Comparison between Water Quality Index (WQI) and biological indices, based on planktonic diatom for water quality assessment in the Dong Nai River, Vietnam
.
Pollution
3
(
2
),
311
323
.
Rothenberger
M. B.
,
Swaffield
T.
,
Calomeni
A. J.
&
Cabrey
C. D.
2014
Multivariate analysis of water quality and plankton assemblages in an urban estuary
.
Estuaries and Coasts
37
(
3
),
695
711
.
Sharma
S.
&
Chhipa
R. C.
2016
Seasonal variations of ground water quality and its agglomerates by water quality index
.
Global Journal of Environmental Science and Management
2
(
1
),
79
86
.
Wu
J.
,
Cheng
S. P.
,
Li
Z.
,
Guo
W. J.
,
Zhong
F.
&
Yin
D. Q.
2013
Case study on rehabilitation of a polluted urban water body in Yangtze River Basin
.
Environmental Science and Pollution Research
20
(
10
),
7038
7045
.
Ye
L. L.
,
Pan
C. R.
,
Zhang
Z. Y.
,
Zheng
Z. X.
,
Liu
J. J.
&
Wang
J. Q.
2006
Characteristics of N forms in Wabu lake sediments and effects of environmental factors on NH4+-N release
.
Journal of Agro-Environment Science
25
,
1333
1336
.
Zhang
W. L.
,
Wang
C.
,
Li
Y.
,
Wang
P. F.
,
Wang
Q.
&
Wang
D. W.
2014
Seeking sustainability: multiobjective evolutionary optimization for urban wastewater reuse in China
.
Environmental Science & Technology
48
(
2
),
1094
1102
.

Supplementary data