Assessing river water quality using water quality index in Lake Taihu Basin, China

Assessing river water quality using water quality index in Lake Taihu Basin, China

Science of the Total Environment 612 (2018) 914–922 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www...

2MB Sizes 0 Downloads 7 Views

Science of the Total Environment 612 (2018) 914–922

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Assessing river water quality using water quality index in Lake Taihu Basin, China Zhaoshi Wu ⁎, Xiaolong Wang, Yuwei Chen, Yongjiu Cai ⁎, Jiancai Deng State Key Laboratory of Lake Science and Environment, Nanjing Institute of Geography and Limnology, Chinese Academy of Sciences, Nanjing 210008, PR China

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• We assessed water quality and its spatial variations in rivers of Lake Taihu Basin. • The water quality was considered as generally “moderate” in this basin. • Significant difference was observed among the 6 river systems. • NH 4 -N, COD Mn , NO 3 -N, DO, and tur are the most effective water quality parameters. • Weighs should be fully considered when using the minimum water quality index method.

a r t i c l e

i n f o

Article history: Received 4 June 2017 Received in revised form 14 August 2017 Accepted 29 August 2017 Available online xxxx Editor: D. Barcelo Keywords: Water quality assessment WQI WQImin River Lake Taihu Basin Yangtze River Delta

a b s t r a c t Lake Taihu Basin, one of the most developed regions in China, has received considerable attention due to its severe pollution. Our study provides a clear understanding of the water quality in the rivers of Lake Taihu Basin based on basin-scale monitoring and a water quality index (WQI) method. From September 2014 to January 2016, four samplings across four seasons were conducted at 96 sites along main rivers. Fifteen parameters, including water temperature, pH, dissolved oxygen (DO), conductivity, turbidity (tur), permanganate index (CODMn), total nitrogen, total phosphorus, ammonium (NH4-N), nitrite, nitrate (NO3-N), calcium, magnesium, chloride, and sulfate, were measured to calculate the WQI. The average WQI value during our study period was 59.33; consequently, the water quality was considered as generally “moderate”. Significant differences in WQI values were detected among the 6 river systems, with better water quality in the Tiaoxi and Nanhe systems. The water quality presented distinct seasonal variation, with the highest WQI values in autumn, followed by spring and summer, and the lowest values in winter. The minimum WQI (WQImin), which was developed based on a stepwise linear regression analysis, consisted of five parameters: NH4-N, CODMn, NO3-N, DO, and tur. The model exhibited excellent performance in representing the water quality in Lake Taihu Basin, especially when weights were fully considered. Our results are beneficial for water quality management and could be used for rapid and low-cost water quality evaluation in Lake Taihu Basin. Additionally, we suggest that weights of environmental parameters should be fully considered in water quality assessments when using the WQImin method. © 2017 Elsevier B.V. All rights reserved.

1. Introduction ⁎ Corresponding authors. E-mail addresses: [email protected] (Z. Wu), [email protected] (X. Wang), [email protected] (Y. Chen), [email protected] (Y. Cai), [email protected] (J. Deng).

http://dx.doi.org/10.1016/j.scitotenv.2017.08.293 0048-9697/© 2017 Elsevier B.V. All rights reserved.

Adequate amounts of suitable quality water resources provide a precondition for economic development and ecological integrity. Numerous stresses influence water quality, such as natural processes

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

(e.g., weathering, precipitation, soil erosion, etc.), anthropogenic activities (e.g., agricultural, urban and industrial activities) and the increased utilization of water resources (Carpenter et al., 1998; Singh et al., 2005; Todd et al., 2012). Due to the multifaceted effects noted above, water quality deterioration has become a serious issue worldwide. Notably, freshwater resources may become scarce in the future, which would threaten water resource use, especially for drinking water and economic development (Cheng et al., 2009; Vorosmarty et al., 2010). According to a study conducted by the World Health Organization (WHO, 2008), approximately 1.1 billion people worldwide do not have access to a reliable source of drinking water. Rivers provide the main water resources for domestic, industrial, and irrigational purposes; however, they are easily polluted because of their critical roles in transporting municipal and industrial pollution and runoff from agricultural land (Singh et al., 2005). Because of their pivotal roles in ecological and human health and economic development, it is essential to prevent and control declining water quality in rivers. Therefore, reliable information regarding water quality variations must be collected for effective management; this has already been conducted in many countries and regions (Astel et al., 2006; Behmel et al., 2016; Romero et al., 2016). The highly heterogeneous water quality variations in rivers should be analyzed at a sufficient spatial scale using hydrochemical monitoring methods (Qadir et al., 2008; Singh et al., 2005). Additionally, water quality evaluation is important for pollution control and resource management. Specifically, the status and trends of water quality can be determined by evaluation, and water quality assessment is critical for identifying the major contributors to spatial and temporal variations in quality, which can benefit water resource management. Furthermore, based on information from assessments, the public is more likely to implement protective measure to improve the conditions of water bodies. The water quality index (WQI) method has been widely used in water quality assessments of both groundwater and surface water, particularly rivers, and it has played an increasingly important role in water resource management (Debels et al., 2005; Lumb et al., 2011; Mohebbi et al., 2013; Sutadian et al., 2016). Using a method that combined the WQI method and geographic information systems, Sener et al. (2017) assessed spatial variations in water quality in the Aksu River, Turkey. Compared with traditional water quality evaluation, WQI methods combine multiple environmental parameters and effectively convert them into a single value reflecting the status of water quality. Thus, instead of comparing the different assessment results of various parameters, the WQI method is an effective approach to water quality assessment and management and provides integrated information regarding the overall quality. A minimum WQI (WQImin) is of great importance in light of simplification based on crucial environmental parameters that influence water quality. WQImin results have been shown to be highly linearly correlated with those of the WQI (Akkoyunlu and Akiner, 2012; Sanchez et al., 2007), indicating that the WQImin approach is powerful with respect to the rapid determination of WQI. With the use of limited parameters, WQImin is particularly beneficial to decreasing the analytical cost of measuring a large number of environmental parameters, especially in developing countries. Generally, the parameters selected for WQImin calculation should be representative of other environmental parameters and must be easily measured (Pesce and Wunderlin, 2000). In this context, the key parameters used for WQImin calculation vary and depend on the characteristics of water bodies. For example, the use of the dissolved oxygen deficit was suggested for rapid determination of the WQI in watersheds of Las Rozas, Madrid (Spain) (Sanchez et al., 2007). Simoes et al. (2008) proposed a WQImin method based on turbidity, total phosphorus and dissolved oxygen (DO) to evaluate degradation in the Medio Paranapanema Watershed, Sao Paulo State, Brazil. Another version of WQImin was developed by Sun et al. (2016) based on the pH, temperature, total suspended solids, ammonium, and nitrate to evaluate the spatial and temporal variations in the water quality of the Dongjiang River, China.

915

A WQImin based on the mean value of three parameters (turbidity, DO, and either conductivity or dissolved solids) after normalization was originally proposed by Pesce and Wunderlin (2000). The same calculation was adopted by Simoes et al. (2008). In later development, Kocer and Sevgili (2014) defined the calculation of WQImin as being similar to WQI, while weights were partially considered. The use of weights in the calculation of WQImin was also partially considered in recent studies (Avigliano and Schenone, 2016; Naveedullah et al., 2016). By contrast, consistency in the calculation of WQI and WQImin has been observed in a limited number of studies (e.g., Zhao et al., 2013). Therefore, the performance of WQImin with and without considering the weights of environmental parameters needs further evaluation. Lake Taihu Basin, which is one of the most developed regions in China, has received considerable attention due to its economic role, as well as the severe pollution resulting from agricultural, urban and industrial activities in this area. Notably, Lake Taihu, the third largest freshwater lake in China, located in the center of the basin, experiences numerous ecological problems, especially eutrophication and cyanobacterial blooms (Chen et al., 2003; Paerl et al., 2011; Qin et al., 2007). Additionally, rivers, especially inflows, have an important influence on the water quality of the lake. Pollution control and natural resource management in Lake Taihu Basin represent notable challenges for the local government. As an issue of national concern, numerous studies in this basin have focused on physical, chemical (e.g., total nitrogen, total phosphorus, heavy metals, etc.), and biological (e.g., phytoplankton, benthic macroinvertebrates, etc.) parameters, as well as land use influences (Bian et al., 2016; Huang and Gao, 2017; Mu et al., 2015; Wu et al., 2011; Wu et al., 2016). Moreover, some studies have addressed the spatial variations in water quality in Lake Taihu Basin. For example, Li et al. (2013) assessed the surface water quality of the Tiaoxi River based on support vector machine classification models; however, the models could not analyze the water quality characteristics throughout the entire basin. Based on benthic macroinvertebrates, Wu et al. (2011) and Huang et al. (2015) assessed the ecological conditions in Lake Taihu Basin using different indexes; however, the accuracy of these assessments depended on the macroinvertebrate knowledge of professionals, and the methods were time consuming. Furthermore, evaluation results varied when single-factor or different biological indexes were applied (Li et al., 2013; Wu et al., 2011). Therefore, a study based on the whole basin and a more suitable method of water quality assessment are needed in this basin. In this study, the WQI method was applied to assess the water quality and its spatial variations in rivers in Lake Taihu Basin. Our study was based on a data set of 15 parameters measured four times at 96 sampling sites that covered the entire basin. The primary objectives of this study were (1) to determine the water quality status and its spatial variation across the basin and (2) to explore the critical parameters in the development of a WQImin method for simple and cost-effective water quality evaluation. We expect that the performance of WQImin will improve with full consideration of parameter weights. 2. Material and methods 2.1. Study area Lake Taihu Basin (30°7′19″–32°14′56″ N, 119°3′1″–121°54′ 26″ E) is located in the lower portion of the Yangtze River (Fig. 1) and encompasses a watershed area of 36,895 km2. The basin covers parts of Jiangsu and Zhejiang Provinces and Shanghai municipality, which are highly developed and populated areas. The population and population density are 59.20 million and 1600 inhabitants km−2, respectively. With 4.4% of the national population, Lake Taihu Basin contributed to 10.4% of China's gross domestic product (GDP) in 2012 (RMB 5418.8 billion) (Lake Taihu Basin Authority, 2012). The total length of the rivers in Lake Taihu Basin is approximately 120,000 km. N200 rivers are distributed across the basin, and most of them connect to Lake Taihu. Generally, these rivers can be divided into

916

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

Fig. 1. Location of the sampling sites in Lake Taihu Basin, China (TG: Tiaoge river system; NH: Nanhe river system; TX: Tiaoxi river system; YJ: Yanjiang river system; HP: Huangpu river system; HY: Hangzhou Bay and the Yangtze Estuary river system).

6 river systems: the Taoge (TG), Nanhe (NH), Tiaoxi (TX), Yanjiang (YJ), Huangpu (HP), and Hangzhou Bay and the Yangtze Estuary (HY) river systems. TX is located in the northern part of Zhejiang Province, with a basin area of approximately 4579 km2 and a length of 293 km. The average annual rainfall and temperature of TX is 1460 mm and 15.6 °C, respectively. The mainstream length of NH is approximately 50 km. TX and NH are mainly located in the western hilly region of the basin, and the mean elevations of these two river systems are relatively high. TG is mainly located in Changzhou City and part of Zhenjiang and Wuxi City. The digital elevation model (DEM) of Lake Taihu Basin is shown in Fig. S1. TX, NH, and TG flow into Lake Taihu and account for 20%, 25%, and 50%, respectively, of the runoff that enters Lake Taihu. The YJ, HP, and HY systems are outflows of Lake Taihu. YJ comprises the northern rivers that connect to the Yangtze River, and the gates to the Yangtze River are controlled. With an area of approximately 14,000 km2, HP is the main river system in the basin, including most of the plain area. The Huangpu River is the mainstream of HP and is the only river connected to the Yangtze River without gate control. The HY system contains rivers that connect to Hangzhou Bay and the Yangtze Estuary in the east. The land use varies among the 6 river systems (Fig. S2). Forest land is mainly observed in TX and NH, and built-up area dominates in HP and YJ. 2.2. Sample collection and laboratory analysis Ninety-six sampling sites were carefully selected to represent the whole of Lake Taihu Basin, covering the main rivers of all 6 river systems

(Fig. 1). The HP, YJ, TX, HY, TG and NH systems contained 28, 21, 15, 14, 11, and 7 sampling sites, respectively. Geographical locations of sampling sites were recorded using a portable GPS system (GPS 639 sc, China). Four sampling campaigns were conducted in September 2014, March 2015, July 2015, and January 2016, spanning seasonal variations, and four samples were collected at each site. The geographic location and sampling time of each site are listed in the supplementary material (Table S1). Most of the samples were collected under sunny or cloudy weather conditions to minimize the effects of rainfall. Selected environmental parameters, including surface water temperature (T), pH, conductivity (cond), turbidity (tur) and DO were obtained in situ using a Hydrolab Datasonde 5 sensor (USA). The sensors were calibrated prior to sampling. The surface water samples (from a depth of ~ 20 cm (PN-ISO 5667–5, 2003; PN-ISO 5667–6, 2003)) were collected in acidcleaned, 10-L plastic buckets that were rinsed with surface water. Samples used for the determination of ammonium (NH4-N), nitrite (NO2N), nitrate (NO3-N), calcium (Ca), magnesium (Mg), chloride (Cl), and sulfate (SO4) concentrations were filtered using GF/F filters (Whatman, Kent Great Britain). The samples were stored in a refrigerator filled with ice. Because of the considerable area of the basin, each sampling event was completed within 3 weeks, and samples were transported from the field to the laboratory every 3–5 days using cold storage. The concentrations of environmental parameters concentrations, including total nitrogen (TN), total phosphorus (TP), NH4-N, NO2-N, NO3-N, permanganate index (CODMn), Ca, Mg, Cl, and SO4, were measured in the State Key Laboratory of Lake Science and Environment

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

(Nanjing, China). The detailed methods and the instruments used for the water chemistry analyses can be found in the supplementary material (Table S2).

917

the WQImin value with no transformation based on the selected parameters and O is the WQI value according to all parameters. Pearson correlation analysis was performed to detect relationships between WQI and WQImin using the SPSS statistical package for Windows (version 17.0).

2.3. WQI calculations and WQImin establishment 3. Results The WQI calculations used here were proposed by Pesce and Wunderlin (2000). Each environmental parameter was assigned a weight based on its perceived effect on primary health (Table S3). T, pH, cond, tur, DO, TN, NH4-N, NO2-N, NO3-N, Ca, Mg, Cl, and SO4 were used to calculate the WQI, and their measured values were used for normalization. The WQI equation was established as follows: n

WQI ¼

∑i¼1 C i P i n Σⅈ¼1 P i

ð1Þ

where n is the total number of parameters included in the study, Ci is the normalized value of parameter i, and Pi is the weight of parameter i. The minimum value of Pi was 1 (Table S3), and the maximum weight assigned to parameters that affect water quality was 4; these values have been verified in previous publications (Debels et al., 2005; Kocer and Sevgili, 2014; Zhao et al., 2013). The WQI ranges from 0 and 100, with high values representing good water quality conditions. The water quality was classified into five grades based on the WQI scores: excellent (91–100), good (71–90), moderate (51–70), low (26–50), and bad (0–25) (Jonnalagadda and Mhere, 2001). The WQI value of each sampling site was calculated on a seasonal basis and was averaged to determine the final WQI value. To develop a simple and cost-effective water quality evaluation approach using a few critical parameters in Lake Taihu Basin, both nonweighted WQImin-a and weighted WQImin-b were considered here. The traditional non-weighted WQImin-a was calculated according to eq. (2). WQImin-b assigns weights to the critical parameters based on the calculation of WQI expressed in eq. (1). n

WQImin−a ¼

∑i¼1 C i n

ð2Þ

The parameters included in the calculations of WQImin-a and WQImin-b were the same in our study. The selection of these parameters was based on the results of linear regression analysis. The establishment of WQImin in Lake Taihu Basin was divided into two steps: training and testing (Wu et al., 2012). First, 288 samples collected during the first three sampling events (September 2014, March 2015, and July 2015) were used to train WQImin. Then, we tested WQImin using 96 samples collected in January 2016. Several models of WQImin (Table 2) were tested to obtain the most appropriate WQImin model for this basin with optimal suitability and predictive ability. 2.4. Data analysis Significant differences in the data at the spatial and seasonal scales were evaluated with Kruskal–Wallis tests using Paleontological Statistics v2.15 (PAST) software. To explore the key parameters that influence the WQI values and to develop a WQImin expression for Lake Taihu Basin, stepwise multiple linear regressions were performed, including T, pH, cond, tur, DO, TN, NH4-N, NO2-N, NO3-N, TP, CODMn, Ca, Mg, Cl, and SO4. The environmental parameter data used in regression analysis were normalized values, i.e., Ci, excluding the effect of weight in the WQI calculation. All data were log (x + 1) transformed prior to analysis (with the exception of significant analysis) to meet the conditions of normality and homogeneity of variance conditions. The coefficient of determination (R2) was used to assess the goodness of fit of the models. The percentage error (PE) of the model was used to evaluate the predicP tive ability of different WQImin expressions. PE was calculated as PE= j OP −1j  100=n according to Canfield and Bachmann (1981), where P is

3.1. Water characteristics Table 1 provides a statistical summary of the water quality parameters measured in the 6 river systems. Kruskal–Wallis tests showed that most parameters varied significantly among the 6 river systems (with the exception of T and NO3-N). The average pH was highest in NH (8.20), followed by HY, and the lowest value was observed in YJ (7.40). The mean cond values in NH and TX were 361.36 μS cm−1 and 207.29 μS cm− 1, respectively, which were significantly lower than those in the four other river systems; all P values all were b 0.05. The tur value was highest in TG (mean value of 147.47 NTU), and no significant tur difference existed between TG and the other river systems, except TX, which exhibited the lowest average value (84.87 NTU). The highest average DO value was observed in HY (9.41 mg L− 1), while the lowest average value was observed in YJ (7.89 mg L−1). The nutrient concentrations were relatively low in TX and NH. The average values of TN in NH and TX were 3.11 mg L−1 and 3.06 mg L− 1, respectively, and both were significantly lower than those in the four other river systems. This distribution was similar to that of NO2-N among the 6 river systems. The mean concentration of NH4-N was lowest in TX and NH with values of 0.23 mg L− 1 and 0.59 mg L−1, respectively, and the value in TX was significantly lower than the values measured in TG, YJ, HP, and HY (all P values b 0.01). In NH, the mean TP concentration was 0.13 mg L−1, which was significantly different from the concentrations measured in TG and YJ (P values of 0.002 and b 0.001, respectively), and the TP concentration in TX was lower than that in NH. In TX, the mean concentration of CODMn was 2.68 mg L−1. This value was significantly lower than the CODMn concentrations in the five other river systems (all P values b 0.001), which had mean values ranging from 4.18 mg L−1 to 4.74 mg L−1. The Ca concentration did not differ significantly among TG, NH, YJ, HP, and HY, but the mean Ca concentration of 15.78 mg L−1in TX was significantly lower than the values measured in the other river systems (all P values b 0.01). A similar situation was observed for Cl, which had an average concentration of 11.08 mg L−1 in TX and was greater than or equal to 29.75 mg L−1 in the other river systems. Relatively high and low mean concentrations of Mg were observed in HY and TX, with values of 8.57 mg L−1 and 4.01 mg L−1, respectively; however, the Mg concentrations in the other river systems varied over a narrow range (6.10 mg L−1–6.85 mg L−1). The mean concentrations of SO4 in TX and NH were relatively low, with values of 31.80 mg L−1 and 45.19 mg L−1, respectively, which were at least 20 mg L−1 lower than the concentrations in the other river systems. 3.2. Water quality assessment based on the WQI In general, most of the sampling sites (88.54% of the 96 sites) in Lake Taihu Basin were classified as having a “moderate” water quality during the study period (Fig. 2). Only 9 and 1 sampling sites had a “good” and “low” water quality status, respectively. Overall, the water quality of the 6 river systems was classified in “moderate” status, whereas the WQI values differed significantly among these systems (Fig. 3). The highest average WQI value was observed in TX (69.28), followed by NH (64.86). The WQI values in TX and NH were significantly higher than those in TG, YJ, HP, and HY (Fig. 3). With the lowest value of 54.75, the average WQI in YJ was significantly different from that of HP, TX and NH (all P values b0.05). The WQI values in HP and YJ ranged from 35.19 to 80.74 and from 34.44 to

918

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

Table 1 Environmental variables summarized as the mean and standard deviation for the 6 river systems in Lake Taihu Basin (TG: Tiaoge river system; NH: Nanhe river system; TX: Tiaoxi river system; YJ: Yanjiang river system; HP: Huangpu river system; HY: Hangzhou Bay and the Yangtze Estuary river system). Parameter

T (°C) pH Cond (μS cm−1) Tur (NTU) DO (mg L−1) TN (mg L−1) NH4-N (mg L−1) NO3-N (mg L−1) NO2-N (mg L−1) TP (mg L−1) CODMn (mg L−1) Ca (mg L−1) Mg (mg L−1) Cl (mg L−1) SO4 (mg L−1)

River systems

Statistics

TG (n = 44)

NH (n = 28)

TX (n = 60)

YJ (n = 84)

HP (n = 112)

HY (n = 56)

H

P

18.30 ± 10.03 7.71 ± 0.80ab 498.89 ± 202.75a 147.47 ± 168.64a 9.08 ± 2.97ab 4.70 ± 1.64a 1.40 ± 1.19ab 1.45 ± 0.98 0.10 ± 0.08a 0.25 ± 0.16a 4.47 ± 1.37a 26.12 ± 11.93a 6.85 ± 5.17ab 48.23 ± 61.84a 66.43 ± 56.82ab

19.74 ± 10.52 8.20 ± 0.45a 361.36 ± 125.27b 129.41 ± 152.46ab 9.04 ± 2.35ab 3.11 ± 1.11b 0.59 ± 0.50cd 1.10 ± 1.04 0.04 ± 0.04b 0.13 ± 0.08bc 4.60 ± 1.43a 23.68 ± 9.85a 6.19 ± 4.23ab 29.75 ± 20.99a 45.19 ± 31.01bc

19.56 ± 9.49 8.05 ± 0.54a 207.29 ± 93.31c 84.87 ± 83.79b 8.79 ± 2.84ab 3.06 ± 0.98b 0.23 ± 0.24d 1.29 ± 1.11 0.04 ± 0.03b 0.10 ± 0.08c 2.68 ± 0.98b 15.78 ± 9.25b 4.01 ± 3.29b 11.08 ± 7.71b 31.80 ± 22.97c

18.64 ± 9.45 7.40 ± 0.84b 551.90 ± 176.54a 107.01 ± 88.89ab 7.89 ± 3.54b 4.90 ± 2.01a 1.63 ± 1.42a 1.26 ± 1.01 0.10 ± 0.12a 0.27 ± 0.16a 4.48 ± 1.38a 29.54 ± 14.73a 6.10 ± 4.51ab 42.60 ± 28.92a 82.25 ± 66.19a

18.88 ± 9.58 7.61 ± 0.72b 498.12 ± 154.66a 104.21 ± 97.09ab 8.62 ± 2.92ab 4.23 ± 1.87a 1.11 ± 1.47bc 1.33 ± 1.04 0.11 ± 0.12a 0.22 ± 0.17ab 4.18 ± 1.23a 24.52 ± 11.15a 6.35 ± 5.71ab 36.62 ± 22.38a 75.54 ± 68.46a

19.69 ± 9.10 8.15 ± 0.68a 591.18 ± 280.36a 109.16 ± 125.18ab 9.41 ± 2.75a 4.89 ± 1.65a 1.13 ± 1.09ac 1.59 ± 1.16 0.13 ± 0.11a 0.23 ± 0.16ab 4.74 ± 1.25a 25.29 ± 15.04a 8.57 ± 7.37a 59.51 ± 63.67a 66.52 ± 63.54ab

1.40 49.07 144.7 11.57 14.02 65.55 105.5 6.77 77.45 71.45 81.49 44.48 17.65 109.6 56.64

0.92 b0.001 b0.001 0.041 0.016 b0.001 b0.001 0.24 b0.001 b0.001 b0.001 b0.001 0.0034 b0.001 b0.001

Mean values with different letters (a, b, c, and d) are significantly different (P b 0.05).

77.60, respectively. Specifically, no sampling site had WQI values lower than 51 in TX and NH based on all four samplings, indicating that the water quality in these two river systems was “moderate” or better (Fig. S3). In TX, nearly 40% of sites were classified as “good” among the 60 samplings, and the proportion of “good” NH sites was 17.86%. However, 32.14% and 20.45% of sites were classified as “low” quality

in YJ and TG, respectively. Additionally, no WQI value above 70, i.e., no “good” sites, was observed in TG or HY. The water quality exhibited distinct seasonal variation (all P values b0.001, Fig. 4) with the highest WQI values in autumn (mean value of 65.80), followed by spring (59.00), summer (58.87), and winter (53.67). Moreover, similar seasonal patterns of WQI values were

Fig. 2. Spatial distribution of water quality index (WQI) values in Lake Taihu Basin based on the average values from 2014 to 2016.

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

919

that the WQImin-b1 and WQImin-b4 were not suitable for representing WQI values in Lake Taihu Basin. Additionally, the WQImin-b3 model presented the lowest PE (5.67%, Table 2), indicating its potential as the best WQImin model for WQI in this basin. In the testing data set, WQImin-b3 exhibited excellent performance and a significant and close correlation with WQI, with P and R2 values of b 0.001 and 0.90, respectively (Fig. 6). The PE of WQImin-b3 was 8.57%, which is relatively low. The performance of WQImin-a3, without considering weights in the calculation, was also examined. Based on the R2 (0.77) and PE (14.93%) values, the a3 model performance considerably poorer than the b3 model. Additionally, the 95% confidence limit was slightly wider for WQImin-a3 than that for WQImin-b3. 4. Discussion 4.1. Physicochemical parameters and status Fig. 3. Differences in the distributions of WQI values in the 6 river systems of Lake Taihu Basin. Different letters (a, b, and c) indicated significant difference (P b 0.01) between mean values. For abbreviations of river systems, see Fig. 1.

observed in all 6 river systems (Fig. S4). According to the WQI classification, the water quality in all four seasons was rated as “moderate”. 3.3. Training and testing of WQImin Stepwise multiple linear regressions showed that NH4-N made the largest contribution to WQI values based on the training data set (Model 1, R2 = 0.539, P b 0.001, Table S3). CODMn, NO3-N, and DO were selected sequentially and considerably increased the R2 value of the models (Models 2–4, Table S3). Additionally, NO2-N and tur also slightly enhanced the performance of the models (Models 5–6, Table S3). Hence, NH4-N, CODMn, NO3-N, and DO were established as the basic and critical parameters in the training of WQImin. Additionally, the effects of NO2-N and tur on the performance of WQImin were evaluated, based on the R2 and PE. Generally, there were small changes in the R2 values of all models (0.91–0.97) regardless of whether NO2-N, tur and the parameter weights were considered. However, the PE values of the weighted models (WQImin-b) were generally lower than those of the non-weighted models (WQImin-a, Table 2), indicating that calculations that included weights could better explain the variation in the WQI. Furthermore, significant differences were observed between WQI and WQImin-b1/WQImin-b4 in the training data set (Fig. 5), indicating

The distribution of water quality parameters in Lake Taihu Basin was generally consistent with the results of our previous study (Wu et al., 2011) and was comparable to that of other studies. The concentrations of pH, tur, and NO3-N in HP were slightly higher than those observed by Zhang et al. (2017) in HP based on the classification in our study. Ion (Ca, Mg, Cl, and SO4) and TP concentrations decreased to different extents according the findings of Mu et al. (2015) based on data from 2010. This result reflects the pollution control and management efforts over the past 5 years. However, TN was 65.87% higher in our study, with a mean value of 4.18 mg L−1, and Li et al. (2013) stated that TN should be a major reduction target in the environmental management of TX. Our study indicated that the water quality in Lake Taihu Basin was generally classified as “moderate”. Several studies have addressed the water quality in parts of this basin. Based on fuzzy synthetic evaluation, water quality mainly ranged from “bad” to “moderate” in Lake Taihu Basin in 2001 and 2002, reflecting grades lower than those calculated in the study of Zhao et al. (2011). This difference may have occurred for several reasons. First, TX and most of NH were not included in the study of Zhao et al. (2011), and these areas are largely classified as “good” water quality areas. Additionally, the effect of over 10 years of management was noteworthy. Based on an assessment of macrozoobenthos, the “moderate” water quality rating in Lake Taihu Basin is accurate. Combining BPI and Wright indexes, Wu et al. (2011) suggested that the water quality in this basin was generally moderate. Using multimetric indexes, Huang et al. (2015) suggested that the general water quality of rivers and streams in Lake Taihu Basin was less than “good”. 4.2. The difference in water quality among river systems

Fig. 4. Seasonal variation of water quality index (WQI) in Lake Taihu Basin (2014–2016). Different letters (a, b, and c) indicated significant difference (P b 0.01) between mean values.

Although the overall rating in our study was “moderate”, the water quality exhibited a significant difference among the 6 river systems. Notably, the water quality levels in TX and NH were relatively better than those in the rivers outflowing from Lake Taihu (i.e., YJ, HP, and HY) and TG. The water qualities in all river systems were generally in accordance with the variations in the environmental parameters. All environmental parameters in TX except T, pH, and NO3-N were the lowest in our study, and these main parameters exhibited the second lowest values in NH. Conversely, the concentrations of TN, NH4-N, TP, Ca, and SO4 were the highest in YJ. Our evaluation results are comparable to those of previous studies that focused on parts of Lake Taihu Basin. The superior water quality in TX was also manifested by individual parameters, such as DO, CODMn, NH4-N and TP; however, this trend was not observed in TN (Li et al., 2013). The area classified as “bad” was located in YJ, while a few sites with “good” water quality were observed in HP, NH, and YJ according to the classification used in the study of Zhao et al. (2011). Moreover, the water quality was generally the poorest in YJ and the best in TX, which was consistent with the result of bioassessment based on macrozoobenthos assemblages (Wu et al., 2011).

920

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

Table 2 Linear models explaining WQI (lg(WQI + 1)) based on the training data set (n = 288). The model results from a stepwise selection procedure using all 15 parameters examined in our study after normalization. Parameters included

NH4-N, CODMn, NO3-N, DO NH4-N, CODMn, NO3-N, DO, NO2-N NH4-N, CODMn, NO3-N, DO, tur NH4-N, CODMn, NO3-N, DO, NO2-N, tur

WQImin-a (non-weighted)

WQImin-b (weighted)

Model

R2

P

PE

Model

R2

P

PE

a1 a2 a3 a4

0.92 0.94 0.94 0.96

b0.001 b0.001 b0.001 b0.001

7.69% 9.66% 10.53% 17.33%

b1 b2 b3 b4

0.91 0.95 0.94 0.97

b0.001 b0.001 b0.001 b0.001

10.30% 6.92% 5.67% 9.46%

Huang et al. (2015) stated that the environmental status in the western hilly area was better than that in the eastern plain. Additionally, Chen et al. (2013) reported a relatively low water quality in TG. Anthropogenic influences and land use are most likely responsible for variations in the water quality of this basin. TX and NH are located in a hilly area, mainly covered by forest, especially TX. With a relatively low level of disturbance from human activities, TX exhibited better water quality than the other river systems. However, water quality degradation occurs from upstream to downstream regions in TX, which is affected by point source pollution from industry and domestic sewage and non-point source pollution from agriculture (Xie et al., 2015). Anthropogenic disturbances are high in NH and TG, especially TG. As a result of Lake Ge (146.9 km2) and Lake Changdang (89.0 km2), the population density is high in TG, and the water quality has deteriorated in this system, especially in typical rivers, such as from Class IV to V in the Taige Canal and Caoqiao River. Additionally, water pollution in TG can be linked to livestock, industrial wastewater and domestic sewage (Zhang et al., 2011). In HP, YJ and HY, which are relatively centrally located in Lake Taihu Basin, the water quality is vulnerable to human and abiotic stresses, as these areas offer convenient transportation and abundant resources. Furthermore, land use change plays an important role in affecting water quality. Increased runoff and impaired water quality are observed when land use changes from natural to urban (Liu et al., 2017; Putro et al., 2016). Land use, which is dominated by built-up areas in HP, YJ, and HY, has notably increased in recent periods (Gao et al., 2010). This change may contribute to the relatively poor water quality in HP, YJ, and HY. Additionally, connectivity to the Yangtze River and Hangzhou Bay is controlled by floodgates, which increase the water retention time and are detrimental to the discharge of pollution. Thus, these factors are responsible for the relatively low WQI values in YJ and HY.

Fig. 5. Comparison of WQImin-b and WQI in Lake Taihu Basin based on the training data set (For the parameters used in different WQImin models, see Table 2). Different letters (a, b, and c) indicated significant difference (P b 0.01) between mean values.

4.3. Crucial parameter selection and the role of weights in the development of WQImin The WQImin proposed in our study consists of five parameters: NH4N, CODMn, NO3-N, DO, and tur, and exhibited excellent performance in reflecting the water quality in Lake Taihu Basin. WQImin plays an important role in the simplification of WQI, which is beneficial for rapid water quality assessment at a relatively low cost. The parameters selected for WQImin calculation should be representative of other environmental parameters and must be easily measured (Pesce and Wunderlin, 2000). According to the stepwise multiple linear regression analysis, the models containing NH4-N, CODMn, NO3-N, DO, and tur were responsible for most of the variance observed in the water quality data in this basin (R2 N 0.88, P b 0.001). NH4-N was the first parameter expressed in the linear regression and contributed the most to the WQI in the training group. As an indicator of organic pollution, CODMn had the second highest explanatory power with respect to WQI in Lake Taihu Basin, followed by NO3-N and DO. Although NO2-N appeared in the WQI model before tur, tur yielded better results in the WQImin model of Lake Taihu Basin. Furthermore, the five parameters selected in our study meet the requirement of ease of measurement and can easily be measured either by automatic or manual monitoring, which is beneficial for the effective evaluation of the water quality in Lake Taihu Basin. The parameters selected in our study generally reflected the environmental conditions of Lake Taihu Basin. Previous studies conducted in this basin are beneficial in assessing the selection of these five parameters in WQImin development. Notably, nitrogen pollution is severe in Lake Taihu Basin (Li et al., 2013; Wang et al., 2007; Xie et al., 2007; Zhang et al., 2011), and NH4-N and NO3-N accounted for a large proportion of dissolved nitrogen according to our study. Xu et al. (2009) identified three principal components (PCs) in a principal component analysis. PC1 and PC2 were associated with strong positive loadings of NH4-N (0.973) and NO3-N (0.912), respectively, which explained 61.61% of the total variance in the water quality in the northwestern portion of Lake Taihu Basin. Another two PCs were extracted by Zhao et al. (2011) over a wider spatial scale in this basin, excluding TX and HY. The results illustrate the differences in the concentrations of NH4N and CODMn and highlight their critical roles in affecting the water quality in Lake Taihu Basin (Zhao et al., 2011). The parameters selected in this study also played a key role in the establishment of WQImin models in other areas. Kocer and Sevgili (2014) suggested that NH4-N was an essential parameter in the assessment of the water quality of Esen Stream (Turkey), which was subjected to trout farm effluent. Combined with NH4-N, NO3-N played a crucial role in WQImin model development in the Chililan River (central Chile) (Debels et al., 2005) and the Dongjiang River (China) (Sun et al., 2016). As an indicator of organic contamination, COD was one of the two most important water quality parameters in the determination of WQI in the Aksu River (Sener et al., 2017), which was also highlighted by Debels et al. (2005). DO is a crucial parameter for aquatic life, and tur is associated with parameters reflecting the underwater light condition as well as bacteriological pollution (Debels et al., 2005). As typical parameters that reflect the water quality, DO and tur have been

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

921

Fig. 6. Relationship between WQI and WQImin-a3 and WQImin-b3 based on the testing data sets and considering the following parameters: ammonium (NH4-N), permanganate index (CODMn), nitrate (NO3-N), dissolved oxygen (DO), and turbidity (tur).

commonly used in WQImin development, including the original WQImin studies (Kannel et al., 2007; Pesce and Wunderlin, 2000; Sanchez et al., 2007; Simoes et al., 2008). The performance of WQImin was improved by fully considering parameter weights in the assessment. Weights, which are commonly considered in WQI calculations, were not considered in the original calculation of WQImin (Pesce and Wunderlin, 2000; Simoes et al., 2008). In later development, some differences remined in the calculation of WQImin, and weights were partially considered in most studies (Avigliano and Schenone, 2016; Naveedullah et al., 2016; Kocer and Sevgili, 2014), while a few studies adopted the same calculation as that used for WQI. In this study, we calculated the WQImin in two ways, i.e., WQImin-a and WQImin-b, and compared the model performance with and without weights based on the same parameters. According to the R2 and PE values of the models, the correlation between WQI and WQImin improved with addition of the weights. Furthermore, through partial consideration of weights, the maximum value of WQImin may exceed 100. As a result, WQImin with partial consideration is probably unsuitable for use in water quality classification. Therefore, we suggest that the calculation method with weights, i.e., the same calculation used for WQI in formula (1), is more appropriate for determining WQImin and for reflecting the water quality, not only in Lake Taihu Basin, but also in water bodies worldwide. 4.4. Uncertainty analysis Uncertainty is inevitable in water quality assessment and is associated with sampling precision, normalization factors, and predictions of WQI in our study. Inherent uncertainty in the dataset is derived from the monitoring process. Certain parameters, such as DO and temperature, are greatly influenced by weather conditions and sampling moments (Debels et al., 2005). Compared with temperature, DO is more important in aquatic ecosystems and is assigned a relatively high weight (4) in water quality assessment. The concentration of DO will vary at different moments of the day and under different weather conditions. Most of our samples were conducted under sunny or cloudy weather conditions to minimize the effects of rainfall. However, the samplings events could not be conducted according to a strict time schedule given the considerable area of the basin. The most affected parameters in the determination of WQI in Lake Taihu Basin were NH4-N, CODMn, and NO3-N. Observed DO and tur were the second most critical parameters in our study area, which may have reduced the noise existing in DO measurement. The normalization factors play an important role in the determination of WQI. The normalization factors of our parameters were mainly

adopted from Pesce and Wunderlin (2000) and Kannel et al. (2007) and have been verified in previous publications (Kocer and Sevgili, 2014; Zhao et al., 2013). Additionally, we took into account the background conditions of China, and we modified the normalization factors of several parameters (i.e., TN, TP, and CODMn) based on the surface water quality standard (GB3838-2002) in China. With regard to the prediction of WQI from WQImin, weighted and non-weighted calculations were both analyzed. The 95% confidence limit was slightly wider for WQImin-a3 than for WQImin-b3 based on the testing dataset, indicating that WQImin considered weights exhibited lower contributions to uncertainty in the prediction of WQI. 5. Conclusions In this study, WQI method was applied to assess the water quality in the rivers of Lake Taihu Basin. The results showed that most parameters varied significantly among the 6 river systems (with the exception of water temperature and NO3-N), and most environmental parameters were the lowest in TX. The water quality was generally “moderate” in this basin according to the WQI classification during our study period. The water quality varied considerably among the 6 river systems, and water quality in the TX and NH rivers was better than that in rivers outflowing from Lake Taihu (i.e., YJ, HP, and HY) and rivers in TG. Moreover, the water quality presented distinct seasonal variation, with the highest WQI values in autumn, followed by spring and summer, and the lowest values in winter. Similar seasonal patterns of WQI values were observed in all 6 river systems. Anthropogenic influences and land use were most likely responsible for the observed spatial variations in water quality. The results obtained in this study are acceptable and comparable to those of previous studies conducted in parts of Lake Taihu Basin. The consistence demonstrates the high applicability of the WQI method of water quality assessment in this basin. Additionally, the WQImin proposed in our study consisted of five parameters, i.e., NH4-N, CODMn, NO3-N, DO, and tur, and exhibited excellent performance in reflecting the water quality of this basin, especially when weights were fully considered. These parameters selected in the WQImin calculation can be easily measured, which benefits rapid, low-cost water quality evaluation in Lake Taihu Basin. Therefore, our results are important for water resource management in this basin. Acknowledgements We thank Dr. Yongnian Gao for providing the digital elevation model (DEM) and land use of Lake Taihu Basin. We are also grateful for the three anonymous reviewers for their insightful comments and helpful

922

Z. Wu et al. / Science of the Total Environment 612 (2018) 914–922

suggestions, which helped to considerably improve the quality of this manuscript. This study was financially supported by the Major Science and Technology Program for Water Pollution Control and Treatment (2014ZX07101-011) and Basic Work of Science and Technology (2013FY111802). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2017.08.293.

References Akkoyunlu, A., Akiner, M.E., 2012. Pollution evaluation in streams using water quality indices: a case study from Turkey's Sapanca Lake Basin. Ecol. Indic. 18, 501–511. Astel, A., Biziuk, M., Przyjazny, A., Namieśnik, J., 2006. Chemometrics in monitoring spatial and temporal variations in drinking water quality. Water Res. 40, 1706–1716. Avigliano, E., Schenone, N., 2016. Water quality in Atlantic rainforest mountain rivers (South America): quality indices assessment, nutrients distribution, and consumption effect. Environ. Sci. Pollut. Res. 23, 15063–15075. Behmel, S., Damour, M., Ludwig, R., Rodriguez, M.J., 2016. Water quality monitoring strategies – a review and future perspectives. Sci. Total Environ. 571, 1312–1329. Bian, B., Zhou, Y., Fang, B.B., 2016. Distribution of heavy metals and benthic macroinvertebrates: impacts from typical inflow river sediments in the Taihu Basin, China. Ecol. Indic. 69, 348–359. Canfield Jr., D.E., Bachmann, R.W., 1981. Prediction of total phosphorus concentrations, chlorophyll a, and Secchi depths in natural and artificial lakes. Can. J. Fish. Aquat. Sci. 38, 414–423. Carpenter, S.R., Caraco, N.F., Correll, D.L., Howarth, R.W., Sharpley, A.N., Smith, V.H., 1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol. Appl. 8, 559–568. Chen, Y.W., Qin, B.Q., Teubner, K., Dokulil, M.T., 2003. Long-term dynamics of phytoplankton assemblages: microcystis-domination in Lake Taihu, a large shallow lake in China. J. Plankton Res. 25, 445–453. Chen, Q., Xu, D.J., Zhang, X., Tang, Y., 2013. A preliminary benthic index of biotic integrity (B-IBI) for bioassessment of the plain waterway network of Taihu Basin. Res. Environ. Sci. 26, 1301–1308 (abstract in English). Cheng, H., Hu, Y., Zhao, J., 2009. Meeting China's water shortage crisis: current practices and challenges. Environ. Sci. Technol. 43 (2), 240–244. Debels, P., Fıgueroa, R., Urrutia, R., Barra, R., Niell, X., 2005. Evaluation of water quality in the Chillán river (Central Chile) using physicochemical parameters and a modified water quality index. Environ. Monit. Assess. 110, 301–322. Gao, Y.N., Gao, J.F., Xu, Y., 2010. Response of landscape ecological risk to land use change in level aquatic eco-functional regions in Taihu Lake watershed. J. Nat. Resour. 25, 1088–1096 (abstract in English). Huang, J.C., Gao, J.F., 2017. An improved ensemble Kalman filter for optimizing parameters in a coupled phosphorus model for lowland polders in Lake Taihu Basin, China. Ecol. Indic. 357, 14–22. Huang, Q., Gao, J.F., Cai, Y.J., Yin, H.B., Gao, Y.N., Zhao, J.H., Liu, L.Z., Huang, J.C., 2015. Development and application of benthic macroinvertebrate-based multimetric indices for the assessment of streams and rivers in the Taihu Basin, China. Ecol. Indic. 48, 649–659. Jonnalagadda, S.B., Mhere, G., 2001. Water quality of the Odzi River in the eastern highlands of Zimbabwe. Water Res. 35, 2371–2376. Kannel, P.R., Lee, S., Lee, Y.S., Kanel, S.R., Khan, S.P., 2007. Application of water quality indices and dissolved oxygen as indicators for river water classification and urban impact assessment. Environ. Monit. Assess. 132, 93–110. Kocer, M.A.T., Sevgili, H., 2014. Parameters selection for water quality index in the assessment of the environmental impacts of land-based trout farms. Ecol. Indic. 36, 672–681. Lake Taihu Basin Authority, 2012. Lake Taihu Basin and Southeast Rivers Water Resource Bulletin (2012). Lake Taihu Basin Authority, Shanghai, pp. 1–10 (in Chinese). Li, W., Yang, M.Y., Liang, Z.W., Zhu, Y., Mao, W., Shi, J.Y., Chen, Y.X., 2013. Assessment for surface water quality in Lake Taihu Tiaoxi River Basin China based on support vector machine. Stoch. Env. Res. Risk A. 27, 1861–1870. Liu, Y., Engel, B.A., Collingsworth, P.D., Pijanowski, B.C., 2017. Optimal implementation of green infrastructure practices to minimize influences of land use change and climate change on hydrology and water quality: case study in spy Run Creek watershed, Indiana. Sci. Total Environ. 601-602, 1400–1411. Lumb, A., Sharma, T.C., Bibeault, J.F., 2011. A review of genesis and evolution of water quality index (WQI) and some future directions. Water Qual. Expos. Hea. 3, 11–24. Mohebbi, M.R., Saeedi, R., Montazeri, A., Vaghefi, K.A., Labbafi, S., Oktaie, S., Abtahi, M., Mohagheghian, A., 2013. Assessment of water quality in groundwater resources of Iran using a modified drinking water quality index (DWQI). Ecol. Indic. 30, 28–34.

Mu, X.Y., Brower, J.C., Siegel, D.I., Fiorentino, A.J., An, S.Q., Cai, Y., Xu, D.L., Jiang, H., 2015. Using integrated multivariate statistics to assess the hydrochemistry of surface water quality, Lake Taihu basin, China. J. Limnol. 74, 234–247. Naveedullah, N., Hashmi, M.Z., Yu, C.N., Shen, C.F., Muhammad, N., Shen, H., Chen, Y.X., 2016. Water quality characterization of the Siling Reservoir (Zhejiang, China) using water quality index. Clean: Soil, Air, Water 44, 553–562. Paerl, H.W., Xu, H., McCarthy, M.J., Zhu, G., Qin, B., Li, Y., Gardner, W.S., 2011. Controlling harmful cyanobacterial blooms in a hyper-eutrophic lake (Lake Taihu, China): the need for a dual nutrient (N & P) management strategy. Water Res. 45, 1973–1983. Pesce, S.F., Wunderlin, D.A., 2000. Use of water quality indices to verify the impact of Cordoba City (Argentina) on Suquia River. Water Res. 34, 2915–2926. PN-ISO 5667–5, 2003. Water Quality – Sampling Part 5: Guidance on Sampling of Drinking Water and Water Used for Food and Beverages. PN-ISO 5667–6, 2003. Water Quality – Sampling Part 6: Guidance on Sampling of Rivers and Streams. Putro, B., Kjeldsen, T.R., Hutchins, M.G., Miller, J., 2016. An empirical investigation of climate and land-use effects on water quantity and quality in two urbanising catchments in the southern United Kingdom. Sci. Total Environ. 548-549, 164–172. Qadir, A., Malik, R.N., Husain, S.Z., 2008. Spatio-temporal variations in water quality of Nullah Aik-tributary of the river Chenab, Pakistan. Environ. Monit. Assess. 140, 43–59. Qin, B.Q., Xu, P.Z., Wu, Q.L., Luo, L.C., Zhang, Y.L., 2007. Environmental issues of lake Taihu, China. Hydrobiologia 581, 3–14. Romero, E., Le Gendre, R., Garnier, J., Billen, G., Fisson, C., Silvestre, M., Riou, P., 2016. Longterm water quality in the lower Seine: lessons learned over 4 decades of monitoring. Environ. Sci. Pol. 58, 141–154. Sanchez, E., Colmenarejo, M.F., Vicente, J., Rubio, A., Garcia, M.G., Travieso, L., Borja, R., 2007. Use of the water quality index and dissolved oxygen deficit as simple indicators of watersheds pollution. Ecol. Indic. 7, 315–328. Sener, S., Sener, E., Davraz, A., 2017. Evaluation of water quality using water quality index (WQI) method and GIS in Aksu River (SW-Turkey). Sci. Total Environ. 584-585, 131–144. Simoes, F.D., Moreira, A.B., Bisinoti, M.C., Gimenez, S.M.N., Yabe, M.J.S., 2008. Water quality index as a simple indicator of aquaculture effects on aquatic bodies. Ecol. Indic. 8, 476–484. Singh, K.P., Malik, A., Sinha, S., 2005. Water quality assessment and apportionment of pollution sources of Gomti river (India) using multivariate statistical techniques - a case study. Anal. Chim. Acta 538, 355–374. Sun, W., Xia, C.Y., Xu, M.Y., Guo, J., Sun, G.P., 2016. Application of modified water quality indices as indicators to assess the spatial and temporal trends of water quality in the Dongjiang River. Ecol. Indic. 66, 306–312. Sutadian, A.D., Muttil, N., Yilmaz, A.G., Perera, B.J.C., 2016. Development of river water quality indices-a review. Environ. Monit. Assess. 188. Todd, A.S., Manning, A.H., Verplanck, P.L., Crouch, C., McKnight, D.M., Dunham, R., 2012. Climate-change-driven deterioration of water quality in a mineralized watershed. Environ. Sci. Technol. 46, 9324–9332. Vorosmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Green, P., Glidden, S., Bunn, S.E., Sullivan, C.A., Liermann, C.R., Davies, P.M., 2010. Global threats to human water security and river biodiversity. Nature 467, 555–561. Wang, X.L., Lu, Y.L., Han, J.Y., He, G.Z., Wang, T.Y., 2007. Identification of anthropogenic influences on water quality of rivers in Taihu watershed. J. Environ. Sci. 19, 475–481. WHO, 2008. Water Sanitation and Health (WSH). WHO http://www.who.int/water_sanitation_health/mdg1/en/index.html, Accessed date: 21 February 2008. Wu, Z.S., Cai, Y.J., Chen, Y.W., Shao, X.Y., Gao, J.F., 2011. Assemblage structure investigation of macrozoobenthos and water quality bioassessment of the main river systems in Taihu Basin. J. Lake Sci. 23, 686–694 (Abstract in English). Wu, N.C., Schmalz, B., Fohrer, N., 2012. Development and testing of a phytoplankton index of biotic integrity (P-IBI) for a German lowland river. Ecol. Indic. 13, 158–167. Wu, P., Qin, B., Yu, G., Deng, J., Zhou, J., 2016. Effects of nutrient on algae biomass during summer and winter in inflow rivers of Taihu Basin, China. Water Environ. Res. 88, 665–672. Xie, Y.X., Xiong, Z.Q., Xing, G.X., Sun, G.Q., Zhu, Z.L., 2007. Assessment of nitrogen pollutant sources in surface waters of Taihu Lake region. Pedosphere 17, 200–208. Xie, X.F., Jiang, G.J., Xiao, C., Wu, T., Bian, H.T., Ma, Y., Chen, J.H., 2015. Evaluation of ecosystem health in Western Tiaoxi River watershed based on matter element model. Acta Sci. Circumst. 35, 1250–1258 (abstract in English). Xu, H., Yang, L.Z., Zhao, G.M., Jiao, J.G., Yin, S.X., Liu, Z.P., 2009. Anthropogenic impact on surface water quality in Taihu Lake region, China. Pedosphere 19, 765–778. Zhang, Y., Gao, Y., Zhang, H., Cao, J., Cai, J., Kong, X., 2011. Water pollution control technology and strategy for river-lake systems: a case study in Gehu Lake and Taige Canal. Ecotoxicology 20, 1154–1159. Zhang, Y.C., Ma, R.H., Hu, M.Q., Luo, J.H., Li, J., Liang, Q.C., 2017. Combining citizen science and land use data to identify drivers of eutrophication in the Huangpu River system. Sci. Total Environ. 584–585, 651–664. Zhao, G.J., Gao, J.F., Tian, P., Tian, K., Ni, G.H., 2011. Spatial-temporal characteristics of surface water quality in the Taihu Basin, China. Environ. Earth Sci. 64, 809–819. Zhao, P., Tang, X.Y., Tang, J.L., Wang, C., 2013. Assessing water quality of three gorges reservoir, China, over a five-year period from 2006 to 2011. Water Resour. Manag. 27, 4545–4558.