The quality of the drinking water source reservoirs has always been a research hotspot. However, few have studies focused on the water quality of reservoirs over a relatively long period with time series analysis. In this paper, based on water quality and hydrological data from 2010 to 2020, considering 8 water quality parameters, CCME-WQI with time series analysis was used to explore the interannual and seasonal changes in water quality in the Weishui Reservoir. Furthermore, the main factors affecting water quality were discussed through correlation analysis. The ARIMA model is used to predict water quality in the future. The results show that the water quality was seriously polluted from 2012 to 2013. After 2018, the water quality gradually improved and stabilized. In addition, the water quality is affected by inflow, showing the characteristics of poor water quality in summer and winter. The key parameters affecting water quality are TN and TP, which are almost 2 times higher than the grade II standard of water quality standard. Through the ARIMA model, it is predicted that CCME-WQI is maintained at 80.46 indicating that the water quality will be stable in the future.

  • CCME-WQI with time series analysis was used to explore changes in water quality of the Weishui Reservoir.

  • The water quality is mainly affected by inflow.

  • The key parameters affecting the water quality are TN and TP.

  • The ARIMA model was used to predict the changing trend of the water quality indicating that CCME-WQI is basically maintained at 80.46 and the water quality will be stable in the future.

Graphical Abstract

Graphical Abstract
Graphical Abstract

In the process of social development, reservoirs play an important role in flood control, power generation, water storage, irrigation, and other projects. Even people's daily drinking water cannot be separated from reservoirs (Tian et al. 2018; Yan et al. 2020). In particular, as a drinking water reservoir, the water quality of the reservoir is related to the health of the local people. Once people drink water with quality problems, it will have a huge impact on people's health and cause a range of diseases (Li & Wu 2019). Therefore, it is necessary to understand the local water quality conditions and its water quality change over time, and the different factors impacting water quality.

Many researchers began to study the theory of water quality evaluation as early as the middle of the 20th century (You et al. 2019). There are a large number of methods of water quality evaluation are proposed, e.g. single factor exponent method (Bi et al. 2018), hierarchical cluster analysis (El-Hames et al. 2013; Ye et al. 2017), principal component analysis (Zitko 2006; Praus 2019; Arora & Keshari 2021), fuzzy neural network evaluation (Cheng et al. 2021) and water quality index (Cong et al. 2020; Zotou et al. 2020), etc. Among them, the water quality index (WQI) method is a numerical method (Han et al. 2020). It is widely used, which can reflect the water quality condition for different purposes (Wu et al. 2017b). WQI model is based on an aggregation function that combines the values of the various water quality parameters as a single value to describe the water quality (Nabizadeh et al. 2013).

The first WQI model was proposed by Horton (Horton 1965) and subsequently further developed by many different investigators, such as the NSF index (Deininger 2009), the House index (Uddin et al. 2021), the CCME Index (Lumb et al. 2006), etc. Among them, CCME-WQI has been widely used to evaluate the water quality of various water bodies, such as rivers, lakes, reservoirs (Gao et al. 2016), and groundwater (Venkatramanan et al. 2015), etc. Many researchers compared and analyzed the advantages and disadvantages of CCME-WQI and other water quality analysis methods. Dai et al. (2019) used CCME-WQI and NPI to evaluate and compare the water quality of the tributaries of the Three Gorges reservoir. It was found that the NPI method overemphasized the most serious pollution factors, while the CCME-WQI method can effectively evaluate the overall water quality. In addition, some studies have compared the Water Framework Directive (WFD) with CCME-WQI on the evaluation of the same water body and found that CCME-WQI evaluation is more reliable (Gikas et al. 2020).

In addition, the water quality of drinking water reservoirs has always been a research hotspot (Gu et al. 2014; Qin et al. 2021). Previous studies mainly focused on the water quality evaluation of reservoir tributaries and reservoir pollution sources (Shen et al. 2014; Zhang et al. 2016). Hou et al. (2016) compared the main pollutants with the water quality of different reservoirs by using WQI. Zhao et al. (2020) used the minimum water quality parameter WQImin to study whether the diversion project has an impact on the water quality of Danjiangkou Reservoir. Wang et al. (2020) studied the impact of reservoir water level declining on reservoir change and pollution source identification. However, few studies have focused on the water quality changes of reservoirs over a relatively long period of time. In particular, most reservoirs with low flow velocity, which is easily lead to eutrophication for a long time (Wu et al. 2017a). Therefore, it is necessary to analyze the change in reservoir water quality over a long time. For long-time water quality analysis, the time series analysis in R is undoubtedly a very useful method (Duan et al. 2018). The time series analysis method is mainly used to solve the time series problems with randomness, seasonality, and stationarity (Kim & King 2020). Especially for the water quality in the reservoir, its seasonality is obvious. Through time series analysis, the trend of water quality change can be found easily. Establishing an autoregressive integrated moving average (ARIMA) model can further quantify the water quality of the reservoir in the next few years so that some useful measures can be taken to manage the reservoir and control the water quality in advance.

In this study, the Weishui Reservoir was selected, which is a high-quality water source in Hubei Province. Eight water quality parameters (pH, CODMn, DO, NH4+-N, TP, TN, F, BOD5) were selected to calculate CCME-WQI. The main purposes of this study are (1) to evaluate the water quality changes and differences in Weishui Reservoir from 2010 to 2021 on an interannual scale and seasonal scale, and (2) to evaluate the impact of inflow and water level on the seasonality of reservoir water quality, (3) to determine the key parameters affecting water quality through correlation analysis, and to find out the pollution source causing the water quality change of the reservoir, (4) to understand the water quality change trend in the next few years through the water quality time series data. Based on the above research objectives, it not only evaluates the drinking water safety of the reservoir but also provides a certain basis for the management policy of the reservoir.

Study area

Weishui Reservoir is in the southwest of Songzi City, Hubei Province, China. It is 14.5 km long from east to west and 6 km wide from north to south with a total surface area of 37 km2. Weishui is a tributary of Songxi River, the four water systems of Dongting Lake. It is located at the junction of Southwest Hubei Province and Northwest Hunan Province. As a new water intake source for Songzi City, the Weishui reservoir can solve the water supply of Songzi City and 8 towns along with the water transmission network. It is in complex terrain with many islands and is known as the first artificial earth dam in Asia. Its altitude decreases from west to east (Figure 1). With a total reservoir capacity of 512 million m3, the reservoir plays an important role in irrigation, flood control, power generation, tourism, and water supply. The climate prevailing in the reservoir is hot with rain in summer and cold and dry in winter. The annual average precipitation in the reservoir area is 1,271.3 mm.

Figure 1

Geographic information and sampling point map of Weishui Reservoir.

Figure 1

Geographic information and sampling point map of Weishui Reservoir.

Close modal

Sampling points and data sources

There are two monitoring points selected to monitor the water quality of the reservoir. One monitoring point is located in Shaxiping (111 °24′13.2″, 29 °54′51.7″), which is located at the junction of the Hubei provinces and the Hunan provinces. It is the inflow control station of the Weishui reservoir and is used to monitor the water quality entering the reservoir. Pollutants such as upstream rainwater can enter the reservoir through the Shaxiping monitoring site. While the other point at Dayanzui (111 °34′55.9″, 29 °57′41.3″), is used to monitor and reflect the water quality in the reservoir (Figure 1). From 2010 to 2013, the water quality data was monitored in the reservoir in January, March, May, July, September, and November, while from 2014 to 2020, the water quality data were monitored in the reservoir every month. To meet the investigation needs, 8 water quality parameters were selected to evaluate the water quality. Those selected parameters are ρ(DO), pH, ρCODMn, ρ(BOD5), ρ(TP), ρ(TN), ρ(NH4-N), and ρ(F).

The water samples were collected at 0.5m below the water surface using a 5 L cleaned plexiglass water collector. Water samples of each site were preserved in polyethylene plastic bottles, which were rinsed at least three times with distilled water and then kept at 4 °C in insulation boxes before analyzing water quality parameters. These parameters, such as pH and DO, were investigated using a multiparametric probe (YSI Incorporated, Yellow Springs, Ohio, USA) in the field, and the sensors of the instrument were calibrated before measurement. In the laboratory, CODMn was analyzed by the permanganate titration; BOD5 was measured by a DO monitor and calculated the reduction of DO in the raw water samples after 5 days; TP was determined using the potassium persulfate molybdenum antimony spectrophotometry; TN was analyzed by potassium persulfate oxidation spectrophotometry; NH4+-N was measured by the Nessler's reagent spectrophotometry; the concentrations of inorganic anions F were gauged using ion chromatography. A detailed description of the water quality analysis methods can be found in the book of national standard methods for examining freshwater and wastewater in China (State Environmental Protection Administration 2002). All water samples were analyzed in the water quality laboratory of Jingzhou hydrology and water resources survey bureau, a (China Metrology Accreditation) CMA laboratory in China.

Analysis methods

Water quality index method (CCME-WQI)

To evaluate the water quality of the Weishui Reservoir, the CCME-WQI method was used. The CCME-WQI method is based on the following three elements: scope (F1), frequency (F2), and amplitude (F3). F1, F2, and F3 are calculated as follows:

F1 (Scope)

F1 is used to represent the percentage of indicators exceeding the standard, which represents the percentage of selected parameters that do not meet their respective parameter standard values at least once during the evaluation period. It can be calculated by Equation (1):
(1)

F2 (Frequency)

F2 represents the percentage of monitoring quantity exceeding the standard, which is used to measure how often a water quality objective is not met. It can be calculated by Equation (2):
(2)

F3 (Amplitude)

F3 is the amplitude, measured by how much the objectives are exceeded, and thus represents the amount by which the failed test values do not meet their objectives.

(Guideline value). The following three steps are required to calculate the F3.

Step 1: The number of times by which an individual concentration is greater than (or less than, when the objective is a minimum) the objective is called an excursion. When the test value must not exceed the objective, Equation (3) was used to calculate the excursion. While the test value must be no less than the objective, Equation (4) was used.
(3)
(4)
Step 2: The total amount by which the individual tests are out of compliance (the normalized sum of excursions) is calculated as the following Equation (5):
(5)
The amplitude (F3) is then calculated by an asymptotic function that scales the normalized sum of the excursions from objectives to yield a value between 0 and 100:
(6)
Finally, the CCME-WQI index is calculated as shown in Equation (7):
(7)

According to the above formulas, it can obtain a CCME-WQI value ranging from 0 to 100. Water quality can be divided into five categories based on the CCME-WQI values calculated above, listed in Table 1 below.

Table 1

Classification of water quality status according to the value of the CCME-WQI

CCME-WQI 95–100 80–94 65–79 45–64 0–44 
Water quality status Excellent Good Fair Marginal Poor 
CCME-WQI 95–100 80–94 65–79 45–64 0–44 
Water quality status Excellent Good Fair Marginal Poor 

To calculate the CCME-WQI value, it is crucial to determine the water quality parameters and objectives. Based on all the monitoring data, all heavy metal index concentrations were within the standard range, so the water quality parameters selected in this study were not included in the heavy metal index. The objectives selected the target concentrations from the Environmental Quality Standards for Surface Water in China (GB3838-2002) published by the Ministry of Environmental Protection in China. As an important drinking water source protection area, the water quality management target of Weishui Reservoir must meet Grade II water quality standards for the Environmental Quality Standards for Surface Water (GB3838-2002). The objectives corresponding to the water quality parameters are listed in Table 2.

Table 2

Water quality objectives used in this assessment

Water quality parameterGB3838-2002 standard value of Grade II
1. pH 6∼9 
2. CODMn 4 (mg/L) 
3. DO 6 (mg/L) 
4. NH4+-N 0.5 (mg/L) 
5. TP 0.25 (mg/L) 
6. TN 0.5 (mg/L) 
7. Fluoride 1 (mg/L) 
8. BOD5 3 (mg/L) 
Water quality parameterGB3838-2002 standard value of Grade II
1. pH 6∼9 
2. CODMn 4 (mg/L) 
3. DO 6 (mg/L) 
4. NH4+-N 0.5 (mg/L) 
5. TP 0.25 (mg/L) 
6. TN 0.5 (mg/L) 
7. Fluoride 1 (mg/L) 
8. BOD5 3 (mg/L) 

Time series analysis

Additive decomposition was performed using RStudio 4.1.2. Time series dates between 2010 and 2020 for the value of the CCME-WQI were analyzed using additive decomposition. In the present study, the time series of CCME-WQI were separated into observed value, seasonal, trend, and random noise components. The observed values show the trend of time series changing with time. Seasonal components express the trend of periodic behavior of time series data. For example, time series data can be monthly or annual. The trend conveys that the data changes over time and may show an additive upward trend, a multiplicative upward trend, or a downward trend. The random noise component explains the rest after removing capture trends and seasonal effects. Finally, through the analysis, the ARIMA model is selected to predict the water quality of the reservoir in the future.

For the ARIMA model. Generally, it can be defined by the following Equations (8)–(11):
(8)
where
(9)
(10)
(11)
where is defined as a white noise process; is defined as backward shift operator (lag operator); is defined as the autoregressive polynomial, and is defined as the moving average polynomial. The expression can be abbreviated as ARIMA (p, d, q). where ‘p’ denotes the autoregressive lag, ‘q’ denotes the moving average, and ‘d’ is the order of differentiation. To predict water quality by the ARIMA model. ARIMA model is usually used for stationary time series, so the stationarity of time series should be tested first. If the time series is unstable, the difference is used to make it stable. To determine whether the stationary sequence is worthy of further analysis, the pure randomness test is carried out for the stationary sequence. The second step is to select appropriate parameters according to the autocorrelation function (ACF) and partial autocorrelation function (PACF) diagrams of stationary data. Finally, Akaike's information criterion (AIC) and Bayesian information criterion (BIC) are used to test the fitting effect and select the optimal prediction model.

Time series analysis through the observation and study of the time series, to find the law of its change and development and predict its future trend. ARIMA is an effective model for forecasting issues, which is widely used in the economy, finance, astronomy, meteorology, oceanics, physics, chemistry, medicine, quality control, and many other fields. It has been especially widely applied in social studies. In the past 2020–2021 years, due to the rapid spread of COVID-19, the time series ARIMA model has been applied to analyze the dynamics of epidemic transmission and predicted the number of epidemic victims in a certain period (Khan & Gupta 2020; Chyon et al. 2021). As a time series model, ARIMA describes the data of different random processes and realizes the linear model. It can flexibly construct the optimal model with many time series transformations by using the method of Box-Jenkins. In this paper, the ARIMA model can not only reflect the dynamic changes in water quality in reservoirs in recent years but also predict the water quality status in the next few years, which provides a basis for the management of subsequent reservoirs.

Correlation analysis

SPSS 26.0 was used to analyze the data, calculate the Pearson correlation coefficient between different water quality parameters and CCME-WQI value, and study the linear relationship between them. P<0.05 was the significant level and P<0.01 was the extremely significant level. Through correlation analysis, the critical factors affecting the water quality of the Weishui Reservoir are studied.

The water quality changes at the interannual scale

To intuitively reflect the change in the Weishui Reservoir water quality, mapping the water quality changes by time series from 2010 to 2020 (Figure 2). Ten years ago, the overall water quality of Weishui Reservoir was good with the average value of CCME-WQI being 84. However, the water quality was severely polluted in 2013, and the mean CCME-WQI values this year were 68.16, and the water quality was considered a ‘fair’ grade. The water quality got better in 2014, which may be related to the adjustment of the local ban on reservoir fish culture policy. By 2016, the water quality of the reservoir was polluted again. This is consistent with the water quality and eutrophication status of representative reservoirs in Hubei Province 2013 published by the Department of water resources of Hubei Province. The water quality of Weishui Reservoir was classified as class III degree water in 2012 and 2013, and the eutrophication assessment was medium eutrophication (Hubei Provincial Department of Water Resources 2013). In addition, it is reported that at the end of 2013, a large area of water bloom broke out in the Weishui Reservoir, and the water quality deteriorated due to the impact of aquaculture fertilizer and paper mill sewage (Jingchu Net – Chutian Metropolis Daily 2014). During the period from 2013 to 2015, the water quality gradually improved but did not return to the level before 2011. By 2018, the water quality has gradually stabilized.

Figure 2

Time series for the CCME-WQI to Weishui Reservoir from 2010 to 2020.

Figure 2

Time series for the CCME-WQI to Weishui Reservoir from 2010 to 2020.

Close modal

Seasonal variation in water quality

Additive decomposition is performed on the time series data CCME-WQI to obtain time series observations, trend factors, seasonal factors, and random factors. This is achieved in this study by the stl function in R. For observations, from 2010 to 2020, the CCME-WQI value of Weishui Reservoir decreased significantly twice, in 2013 and 2016, respectively. From the trend item, the movement of reservoir water quality from the highest point to the lowest can be seen, it is a time series of downtrends. The overall water quality trend is deteriorating, but it tends to be stable. This also indicates the instability of the original water quality data. In this study, the data is processed differentially to make it smooth and determine the value of ‘d’. From the seasonal components, water quality has certain seasonal characteristics. random component captures changes in water quality that cannot be explained by trends or seasonal effects (Figure 3).

Figure 3

Decomposition of CCME-WQI data using additive decomposition.

Figure 3

Decomposition of CCME-WQI data using additive decomposition.

Close modal

Overall, the water quality is relatively better in spring and autumn, average CCME-WQI values in spring and autumn were 82.33 and 82.62, respectively. According to the standard of water quality classification, the water quality in spring and autumn was considered a ‘good ‘grade, while the water quality in the summer and winter is relatively worse and rated a ‘Fair’ grade (Figure 4(a). As shown in Figure 4(b), the conclusions are the same. Water quality in different seasons was shown by calculating the proportion of water quality status in four seasons and found that spring and autumn had ‘excellent’ categories, while not measured in summer and winter.

Figure 4

WQI values for different seasons in the Weishui Reservoir (the season was categorized as spring (March-May), summer (June-August), autumn (September-November), and winter (December-February)).

Figure 4

WQI values for different seasons in the Weishui Reservoir (the season was categorized as spring (March-May), summer (June-August), autumn (September-November), and winter (December-February)).

Close modal

Inflow and water level condition

The seasonal variation of water quality is usually related to the inflow of the reservoir. The inflow of the reservoir in different months from 2010 to 2020 is shown in Figure 5. Throughout the inflow changes, the periodic change of inflow is obvious. In July 2020, the inflow has a maximum peak, with a monthly average of 201 m3/s in this year. there is extreme weather of flood and waterlogging, which led to a significant increase in the inflow and water level of the reservoir.

Figure 5

Reservoir inflow and water level change from 2010 to 2020 in different months.

Figure 5

Reservoir inflow and water level change from 2010 to 2020 in different months.

Close modal

From the monthly change of inflow, the month with the highest inflow is July. Taking July as the boundary, the inflow decreases from both sides. The inflow of Weishui Reservoir in summer was the largest compared to the other seasons, and the lowest inflow was in the winter. From 2010 to 2020, the average inflow of the reservoir is 41.4 m3/s in summer and 4.56 m3/s in winter. Similarly, according to the analysis of the water level in Dayanzui, the higher water level still appears in summer. The water level varied between 85 and 95 meters during the study period. It is generally stable at about 90 meters.

According to the water quality analysis, the water quality of the Weishui Reservoir in summer and winter is worse than in the other two seasons. It is not difficult to find that too high or too low an inflow will affect the water quality of the reservoir. Reservoir inflow is affected by rainfall, which is also affected by geography, climate (Wang et al. 2018), topography, and other factors. It is worth noting that some studies have reported that rainfall and climate conditions affecting inflow will have an impact on reservoir dynamics and water quality (Roselli et al. 2009; Razmkhah et al. 2010). In the summer, rain is abundant. Rainwater rushes down to carry a large amount of nitrogen and phosphorus from agricultural non-point source pollution into the reservoir through runoff, which is likely to cause nitrogen and phosphorus in the reservoir to exceed the standard and even cause blooms (Zeng et al. 2019). And some of the pollutants in the rainwater can directly contaminate drinking water (John et al. 2021), causing the nitrogen and phosphorus in the reservoir to exceed the standard, and even causing a bloom. On the other hand studies by Cooper and Tait reported that flow brings a large amount of sediment transport and settles to the bottom of the reservoir, thus forming the water-worked rough bed (Cooper & Tait 2008). While microbes, such as viruses, bacteria, and protozoa, can attach to sediment, causing microbial contamination and deterioration in water quality (Characklis et al. 2005; Krometis et al. 2013). Some studies have suggested that non-point source pollution intensifying runoff from cropland during periods of high flow may be the main cause of poor water quality in summer (Shen et al. 2008). So, the seasonal variation of water quality is closely related to the reservoir inflow, and part of the explanation for the observed phenomenon was that the reservoir inflow is large in summer, and a large amount of nitrogen and phosphorus in agricultural non-point source pollution enters the reservoir through runoff. However, in winter, once the reservoir is polluted, the low inflow of the reservoir makes the reservoir's self-purification capacity decrease, resulting in serious internal pollution of the reservoir. The inflow and water level of the reservoir affect its self-purification capacity. As a result, the self-purification capacity of the reservoir is low in winter. This is also one of the reasons why the water quality is better in spring and autumn, but not in summer and winter.

Key water quality parameters to the CCME-WQI

The changes in water quality parameters

The water quality parameters of Shaxiping (MP1, black line in Figure 6) monitoring point before entering the reservoir and Dayanzui (MP2, red line in Figure 6) in the reservoir from 2010 to 2020 were compared and analyzed by time series. In addition, the statistical description of the water quality parameters of the two stations is shown in Table 3. Among the water quality parameters involved in the water quality evaluation, pH, CODMn, F, and BOD5 meet the Class II of Surface Water Environmental Quality Standard (GB3838-2002). Both BOD5 and CODMn indirectly indicate the degree of organic matter contamination in the water based on the dissolved oxygen consumed in the water. While drinking water is one of the largest sources of fluoride intake, excess fluoride can easily lead to poisoning. For 11 years, these water quality parameters in the reservoir have been within the standard limit, indicating that the water quality is relatively safe. DO is lower than the water quality objectives (Table 2) at least once. TP and NH4+-N of the two monitoring stations exceeded the Class II water standard of surface water. Although TN is not measured at the Shaxiping monitoring point, the concentration of TN in most months of the reservoir exceeded the standard quality by 2–3 times.

Table 3

Statistical description of water quality parameters (mg/L exclude pH) of two stations

Sample pointpHCODMnDONH4+-NTPFTNBOD5
Shaxiping Average 8.2 1.47 8.8 0.16 0.04 0.158 – 1.62 
Min 7.5 0.8 5.7 0.03 0.01 0.03 – 0.2 
Max 8.7 2.6 14.0 0.68 0.21 0.49 – 
SD 0.061 0.133 3.473 0.021 0.001 0.0042 – 0.40 
Dayanzui Average 8.3 1.7 8.2 0.17 0.026 0.16 1.30 1.62 
Min 1.1 5.1 0.01 0.01 0.01 0.33 0.5 
Max 8.7 2.7 12.0 0.53 0.08 0.32 2.43 2.9 
SD 0.030 0.134 2.206 0.017 0.001 0.0038 0.187 0.370 
Sample pointpHCODMnDONH4+-NTPFTNBOD5
Shaxiping Average 8.2 1.47 8.8 0.16 0.04 0.158 – 1.62 
Min 7.5 0.8 5.7 0.03 0.01 0.03 – 0.2 
Max 8.7 2.6 14.0 0.68 0.21 0.49 – 
SD 0.061 0.133 3.473 0.021 0.001 0.0042 – 0.40 
Dayanzui Average 8.3 1.7 8.2 0.17 0.026 0.16 1.30 1.62 
Min 1.1 5.1 0.01 0.01 0.01 0.33 0.5 
Max 8.7 2.7 12.0 0.53 0.08 0.32 2.43 2.9 
SD 0.030 0.134 2.206 0.017 0.001 0.0038 0.187 0.370 
Figure 6

Time series diagram of each water quality parameters in Shaxiping (MP1, black line) and Dayanzui (MP2, red line) from 2010 to 2020: (a) pH value; (b) Potassium permanganate index (CODMn); (c) Dissolved oxygen (DO); (d) ammonia nitrogen (NH4+-N); (e) Total phosphorus (TP); (f) Fluoride (F); (g) Five-day biochemical oxygen demand (BOD5).

Figure 6

Time series diagram of each water quality parameters in Shaxiping (MP1, black line) and Dayanzui (MP2, red line) from 2010 to 2020: (a) pH value; (b) Potassium permanganate index (CODMn); (c) Dissolved oxygen (DO); (d) ammonia nitrogen (NH4+-N); (e) Total phosphorus (TP); (f) Fluoride (F); (g) Five-day biochemical oxygen demand (BOD5).

Close modal

The pH can reflect whether the water quality is suitable for the growth of aquatic organisms, which determines the reproduction of organisms and the water quality status in the water body. Whether its value is too high or too low will directly affect the activity of microorganisms in the water body, resulting in the deterioration of the water body. From the time series plots of each parameter, pH at MP1 and MP2 showed similar trends and were within standard limits during the whole study period (Figure 6(a)), which indicates that the water quality is suitable for the growth of aquatic organisms.

DO is an important indicator to measure the self-purification capacity of water bodies, and the content of dissolved oxygen in water is closely related to water temperature and partial pressure of oxygen in the air. The value of dissolved oxygen was lower than the standard value many times (Figure 6(c)). Minimal dissolved oxygen of MP1 in 2011 was 5.7 mg/L and the average dissolved oxygen 11 years per month was 8.8 mg/L. In January 2013, the minimum dissolved oxygen at the MP2 was 5.1 mg/L. It can be seen that in 2013, the water quality was polluted, and the dissolved oxygen concentration in the reservoir was lower than the surface Class II water standard. At the same time, the COD of the reservoir increased because of the pollution of the water quality of the reservoir in 2013 (Figure 6(b)).

Additionally, NH4+-N is the most harmful form of various types of nitrogen. As a nutrient in the water body, it can provide nutrients for algae growth, control ammonia nitrogen to reduce the load of lake reservoir ammonia nitrogen and total nitrogen, and reduce the probability of eutrophication of water bodies. The concentration of NH4+-N was high from 2013 to 2014, and the water quality deteriorates at this time. It reached the maximum value of 0.68 mg/L in 2012 at MP1. subsequently, the concentration of NH4+-N reached the maximum value in May 2013 at MP2 (Figure 6(d)).

Nitrogen and phosphorus are essential nutrients for biological growth. TN and TP monitoring are crucial for water quality, especially in lakes, reservoirs, bays, and other enclosed waters. Phosphorus enrichment is easy to produce eutrophication, and deterioration of water quality. The TP pollution of the two stations is the most serious. In May 2011, TP reached 0.21 mg/L at MP1, almost 10 times higher than the grade II standard of water quality standard GB3838-2002. The concentration of TP at MP1 is higher than that at MP2 (Figure 6(e)). The annual average concentration of TP at MP1 from 2010 to 2020 varies greatly, varying between 0.033 mg/L and 0.055 mg/L. However, the average total phosphorus concentration fluctuates up and down in the range of 0.05 mg/L, almost twice as much as the Class II water quality standard. In general, the water quality of the reservoir changes due to the changes in the water quality of the inflow monitoring station of shaxiping. Combined with the change in the concentration of each parameter at the two stations, it can be deduced that the water quality of the reservoir is mainly affected by external sources, rather than endogenous pollution.

Correlation analysis

To identify the most important factors affecting CCME-WQI, the relationship between different water quality parameters/inflow (FL) and CCME-WQI in the reservoir was analyzed (Table 4). A significant correlation between the pH, TN, TP, and inflow (FL) and the CCME-WQI is largely found. However, no significant relationship was observed between CCME-WQI and DO or between CCME-WQI and CODMn, with p-values of 0.849 and 0.120, respectively. While the lowest DO values are correlated with high inflow. The results of a Pearson correlation analysis showed that the CCME-WQI was negatively correlated with the TP (R=−0.52, P<0.01) (Figure 7(b)). The highest correlation coefficient between TP and CCME-WQI indicates that TP had the greatest effect on water quality. In addition, it was also closely related between TN and CCME-WQI (R=−0.49, P<0.01) (Figure 7(a)). CCME-WQI is negatively correlated with inflow (Figure 7(c)), while inflow is positively correlated with TN significance (Figure 7(d)), which shows that inflow is an important part of reservoir pollution.

Table 4

Pearson correlation matrix between water quality parameters and the overall CCME-WQI

pHCODMnDONH4+-NTPTNFBOD5inflowWater levelCCME-WQI
pH 0.0043 0.17 0.21* −0.0018 0.15 0.064 0.029 −0.045 −0.22* −0.22* 
CODMn  −0.28 0.16 −0.1 −0.2 −0.16 −0.11 0.13 0.34*** 0.15 
DO   0.085 0.13 −0.089 0.092 −0.053 −0.2* −0.44*** −0.019 
NH4+-N    0.07 −0.17 0.15 −0.08 −0.21* −0.23* −0.035 
TP     −0.069 −0.13 −0.021 0.0023 0.051 −0.52*** 
TN      0.051 0.11 0.34*** 0.29** −0.49*** 
F       0.082 0.0041 −0.12 0.0012 
BOD5        0.059 0.15 −0.14 
inflow         0.52*** −0.2* 
Water level          −0.19 
CCME-WQI           
pHCODMnDONH4+-NTPTNFBOD5inflowWater levelCCME-WQI
pH 0.0043 0.17 0.21* −0.0018 0.15 0.064 0.029 −0.045 −0.22* −0.22* 
CODMn  −0.28 0.16 −0.1 −0.2 −0.16 −0.11 0.13 0.34*** 0.15 
DO   0.085 0.13 −0.089 0.092 −0.053 −0.2* −0.44*** −0.019 
NH4+-N    0.07 −0.17 0.15 −0.08 −0.21* −0.23* −0.035 
TP     −0.069 −0.13 −0.021 0.0023 0.051 −0.52*** 
TN      0.051 0.11 0.34*** 0.29** −0.49*** 
F       0.082 0.0041 −0.12 0.0012 
BOD5        0.059 0.15 −0.14 
inflow         0.52*** −0.2* 
Water level          −0.19 
CCME-WQI           

The ‘*’ and ‘***’ indicate significant correlation at level of 0.05 and 0.01(bilateral), respectively.

Therefore, TP and TN are the two main key water quality parameters to affect the CCME-WQI. The changing trend of total phosphorus concentration at MP2 before 2014 is opposite to that at MP1, but the changing trend is the same after 2014. Since 2014, reservoir pollution has mainly come from external pollution. This may also indicate that TP in the reservoir is affected by the tributaries upstream of the reservoir. Most studies show that the pollution sources of the reservoir generally come from the inflow of tributaries, fish farming (Jia et al. 2015), and sediment release (Gao et al. 2014), especially the reservoir non-point source pollution (Hsieh et al. 2010; Chang & Yu 2020). According to relevant research, Non-point source pollution is the main factor for most reservoirs being polluted (Tong et al. 2021).

The elements nitrogen and phosphorus are the main factors affecting reservoir eutrophication. Like most research results, eutrophication is the main reason threatening the reservoir to provide its services (Huang et al. 2020). It is very important to control the nitrogen and phosphorus flowing into the reservoir. While TN is affected by the inflow of the reservoir. Inflow was correlated with DO, NH4+-N, and TN. The relationship between inflow and other parameters reveals the impact of inflow on the water quality of the Weishui Reservoir. Regulating the inflow plays an important role in improving the water quality of the reservoir. Identifying the impact of reservoir inflow on water quality evaluation results is of great significance to improve the water quality management of Weishui Reservoir.

According to a survey, the livestock and poultry breeding around the reservoir is mainly free range, and the agricultural planting area in the upstream is wide. The reservoir is vulnerable to non-point source pollution, especially domestic non-point source pollution, livestock and poultry breeding pollution, agricultural non-point source pollution, etc.

Water quality prediction using the ARIMA model

In this paper, the water quality of the Weishui Reservoir is mainly affected by inflow, and non-point source pollution is the main source of water quality deterioration. The application of the time series ARIMA model to the prediction of reservoir water quality can evaluate the water quality dynamics of short-term in the future. Compared with other water quality evaluation methods, it is relatively simple and has high economic benefits. R Studio 4.1.2. is used as a tool to design the ARIMA models. Using the tsdiag function in the stats package in R to test and output three parts of information: (1) residual time series diagram; (2) autocorrelation diagram of residual sequence; (3) white noise test chart of residual sequence. Through the white noise diagram of the residual sequence in the third part, we can judge whether the fitting model is significantly valid. The p value is 0.8177, which indicates that model ARIMA (1,1,1) can better fit the data. The results of the water quality evaluation are shown in Figure 8 below. According to the prediction results, the water quality of the reservoir will be stable in the next two years. CCME-WQI is maintained at about 80.46.

Figure 7

Correlations between CCME-WQI and parameters TN, TP, inflow (FL); and Correlation between flow and TN.

Figure 7

Correlations between CCME-WQI and parameters TN, TP, inflow (FL); and Correlation between flow and TN.

Close modal
Figure 8

CCME-WQI sequence fitting and prediction diagram of Weishui Reservoir (In the figure: the solid line is the sequence observation value; the red dotted line and the blue solid line are the model fitting value and prediction value; the dark shadow part is the 80% confidence interval of the prediction value, and the light shadow part is the 95% confidence interval of the prediction value).

Figure 8

CCME-WQI sequence fitting and prediction diagram of Weishui Reservoir (In the figure: the solid line is the sequence observation value; the red dotted line and the blue solid line are the model fitting value and prediction value; the dark shadow part is the 80% confidence interval of the prediction value, and the light shadow part is the 95% confidence interval of the prediction value).

Close modal

The study of time series data has a wide range of applications in the real world. Using water quality data from the past period to predict water quality in the future is very necessary for regional water quality management. Through the time series model, the time series data can be decomposed, which is conducive to the analysis of the overall trend change of the data and captures the seasonal component of the time series. In addition, seasonal components can be culled from time series to achieve seasonal adjustment. Using the ARIMA model to understand the future dynamics of water quality in advance is also conducive to the local government to formulate corresponding policies according to the water quality status.

The water quality of the Weishui Reservoir was severely polluted in 2013, and then gradually recovered in 2018. It was negatively correlated with the inflow of the only major tributary of the reservoir. The key parameters affecting water quality in the reservoir were TN and TP, which were almost 2 times higher than the grade II standard of water quality standard. These indicated that the reservoir is vulnerable to non-point source pollution. Through the ARIMA model, it was predicted that CCME-WQI is maintained at 80.46 indicating that the water quality will be stable in the future. The findings of this study will facilitate the prediction of water quality and better demonstrate the dynamic changes in the water quality of the reservoir. It not only helps to better manage the reservoir but also ensures the safety of drinking reservoirs. In future studies, more models can be explored to improve the accuracy of reservoir water quality predictions and provide technical support for the dynamic monitoring of reservoirs.

We sincerely thank the editors and anonymous reviewers for their outstanding comments.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.

Bi
W.
,
Weng
B.
,
Yuan
Z.
,
Ye
M.
,
Zhang
C.
,
Zhao
Y.
,
Yan
D.
&
Xu
T.
2018
Evolution characteristics of surface water quality due to climate change and LUCC under scenario simulations: a case study in the Luanhe River Basin
.
Int. J. Environ. Res. Public Health
15
(
8
).
http://dx.doi.org/10.3390/ijerph15081724
.
Chang
C. L.
&
Yu
Z. E.
2020
Application of water quality model to analyze pollution hotspots and the impact on reservoir eutrophication
.
Environ. Monit. Assess.
192
(
8
),
495
.
http://dx.doi.org/10.1007/s10661-020-08463-3
.
Characklis
G. W.
,
Dilts
M. J.
,
Simmons
O. D.
3rd
,
Likirdopulos
C. A.
,
Krometis
L. A.
&
Sobsey
M. D.
2005
Microbial partitioning to settleable particles in stormwater
.
Water Res.
39
(
9
),
1773
1782
.
http://dx.doi.org/10.1016/j.watres.2005.03.004
.
Cheng
L.
,
Tan
X.
,
Yao
D.
,
Xu
W.
,
Wu
H.
&
Chen
Y.
2021
A fishery water quality monitoring and prediction evaluation system for floating UAV based on time series
.
Sensors (Basel)
21
(
13
).
http://dx.doi.org/10.3390/s21134451
.
Chyon
F. A.
,
Suman
M. N. H.
,
Fahim
M. R. I.
&
Ahmmed
M. S.
2021
Time series analysis and predicting COVID-19 affected patients by ARIMA model using machine learning
.
J. Virol. Methods
301
,
114433
.
http://dx.doi.org/10.1016/j.jviromet.2021.114433
.
Cong
V. H.
,
Lam
N. T.
,
Nui
N. H.
,
Thao
T. P.
,
Giang
N. T. H.
&
Son
C. T.
2020
Assessment of Cau River water quality assessment using a combination of water quality and pollution indices
.
J. Water Supply Res. Technol. AQUA
69
(
2
),
160
172
.
http://dx.doi.org/10.2166/aqua.2020.122
.
Cooper
J. R.
&
Tait
S. J.
2008
The spatial organisation of time-averaged streamwise velocity and its correlation with the surface topography of water-worked gravel beds
.
Acta Geophys.
56
(
3
),
614
641
.
http://dx.doi.org/10.2478/s11600-008-0023-0
.
Dai, J., Jia, B., Luo, W., Mostafaei, A. & Zhu, S. 2019 Assessing water quality for urban tributaries of the Three Gorges Reservoir, China. Journal of Water Reuse and Desalination 9 (1), 105–114. http://dx.doi.org/10.2166/wrd.2018.010
.
Deininger
R. A.
2009
A water quality index for rivers
.
Water Int.
5
(
3
),
16
21
.
http://dx.doi.org/10.1080/02508068008685872
.
Duan
W.
,
He
B.
,
Chen
Y.
,
Zou
S.
,
Wang
Y.
,
Nover
D.
,
Chen
W.
&
Yang
G.
2018
Identification of long-term trends and seasonality in high-frequency water quality data from the Yangtze River basin, China
.
PLoS One
13
(
2
),
e0188889
.
http://dx.doi.org/10.1371/journal.pone.0188889
.
El-Hames
A. S.
,
Hannachi
A.
,
Al-Ahmadi
M.
&
Al-Amri
N.
2013
Groundwater quality zonation assessment using GIS, EOFs and hierarchical clustering
.
Water Resour. Manage.
27
(
7
),
2465
2481
.
http://dx.doi.org/10.1007/s11269-013-0297-0
.
Gao
B.
,
Wei
X.
,
Zhou
H.
,
Lu
J.
,
Hao
H.
&
Wan
X.
2014
Pollution characteristics and possible sources of seldom monitored trace elements in surface sediments collected from Three Gorges Reservoir, China
.
Sci. World J.
2014
.
170639. http://dx.doi.org/10.1155/2014/170639
.
Gao
Q.
,
Li
Y.
,
Cheng
Q.
,
Yu
M.
,
Hu
B.
,
Wang
Z.
&
Yu
Z.
2016
Analysis and assessment of the nutrients, biochemical indexes and heavy metals in the Three Gorges Reservoir, China, from 2008 to 2013
.
Water Res.
92
,
262
274
.
http://dx.doi.org/10.1016/j.watres.2015.12.055
.
Gikas
G. D.
,
Sylaios
G. K.
,
Tsihrintzis
V. A.
,
Konstantinou
I. K.
,
Albanis
T.
&
Boskidis
I.
2020
Comparative evaluation of river chemical status based on WFD methodology and CCME water quality index
.
Sci. Total Environ.
745
,
140849
.
http://dx.doi.org/10.1016/j.scitotenv.2020.140849
.
Gu
Q.
,
Deng
J.
,
Wang
K.
,
Lin
Y.
,
Li
J.
,
Gan
M.
,
Ma
L.
&
Hong
Y.
2014
Identification and assessment of potential water quality impact factors for drinking-water reservoirs
.
Int. J. Environ. Res. Public Health
11
(
6
),
6069
6084
.
http://dx.doi.org/10.3390/ijerph110606069
.
Han
Q.
,
Tong
R.
,
Sun
W.
,
Zhao
Y.
,
Yu
J.
,
Wang
G.
,
Shrestha
S.
&
Jin
Y.
2020
Anthropogenic influences on the water quality of the Baiyangdian Lake in North China over the last decade
.
Sci. Total Environ.
701
,
134929
.
http://dx.doi.org/10.1016/j.scitotenv.2019.134929
.
Horton
R. K.
1965
An index-number system for rating water quality
. In:
Water Quality Criteria-Stream vs. Effluent Standards, Water Pollut Control Federation
. pp.
292
315
.
http://doi.org/10.2307/25035201.11
.
Hou
W.
,
Sun
S.
,
Wang
M.
,
Li
X.
,
Zhang
N.
,
Xin
X.
,
Sun
L.
,
Li
W.
&
Jia
R.
2016
Assessing water quality of five typical reservoirs in lower reaches of Yellow River, China: using a water quality index method
.
Ecol. Indic.
61
,
309
316
.
http://dx.doi.org/10.1016/j.ecolind.2015.09.030
.
Hsieh
P.-H.
,
Kuo
J.-T.
,
Wu
E. M.-Y.
,
Ciou
S.-K.
&
Liu
W.-C.
2010
Optimal best management practice placement strategies for nonpoint source pollution management in the Fei-Tsui reservoir watershed
.
Environ. Eng. Sci.
27
(
6
),
441
449
.
http://dx.doi.org/10.1089/ees.2010.0003
.
Huang
L.
,
Mo
Y.
,
Wu
Z.
,
Rad
S.
,
Song
X.
,
Zeng
H.
,
Bashir
S.
,
Kang
B.
&
Chen
Z.
2020
Occurrence, distribution, and health risk assessment of quinolone antibiotics in water, sediment, and fish species of Qingshitan reservoir, South China
.
Sci. Rep.
10
(
1
),
15777
.
http://dx.doi.org/10.1038/s41598-020-72324-9
.
Hubei Provincial Department of Water Resources
2013
Hubei Water Resources Bulletin
.
Jia
B.
,
Tang
Y.
,
Tian
L.
,
Franz
L.
,
Alewell
C.
&
Huang
J. H.
2015
Impact of fish farming on phosphorus in reservoir sediments
.
Sci. Rep.
5
,
16617
.
http://dx.doi.org/10.1038/srep16617
.
Jingchu Net – Chutian Metropolis Daily Photo: Water bloom erupts in Songziweishui Reservoir. Available from: https://www.163.com/news/article/9ICASB2600014Q4P.html.
John
C. K.
,
Pu
J. H.
,
Moruzzi
R.
&
Pandey
M.
2021
Health-risk assessment for roof-harvested rainwater via QMRA in Ikorodu area, Lagos, Nigeria
.
J. Water Clim. Change
12
(
6
),
2479
2494
.
http://dx.doi.org/10.2166/wcc.2021.025
.
Khan
F. M.
&
Gupta
R.
2020
ARIMA and NAR based prediction model for time series analysis of COVID-19 cases in India
.
J. Saf. Sci. Resilience
1
(
1
),
12
18
.
http://dx.doi.org/10.1016/j.jnlssr.2020.06.007
.
Kim
T.
&
King
B. R.
2020
Time series prediction using deep echo state networks
.
Neural Comput. Appl.
32
(
23
),
17769
17787
.
http://dx.doi.org/10.1007/s00521-020-04948-x
.
Krometis
L. A.
,
Noble
R. T.
,
Characklis
G. W.
,
Denene Blackwood
A.
&
Sobsey
M. D.
2013
Assessment of E. coli partitioning behavior via both culture-based and qPCR methods
.
Water Sci. Technol.
68
(
6
),
1359
1369
.
http://dx.doi.org/10.2166/wst.2013.363
.
Li
P.
&
Wu
J.
2019
Drinking water quality and public health
.
Exposure Health
11
(
2
),
73
79
.
http://dx.doi.org/10.1007/s12403-019-00299-8
.
Lumb
A.
,
Halliwell
D.
&
Sharma
T.
2006
Application of CCME water quality index to monitor water quality: a case study of the Mackenzie River Basin, Canada
.
Environ. Monit. Assess.
113
(
1–3
),
411
429
.
http://dx.doi.org/10.1007/s10661-005-9092-6
.
Nabizadeh
R.
,
Valadi Amin
M.
,
Alimohammadi
M.
,
Naddafi
K.
,
Mahvi
A. H.
&
Yousefzadeh
S.
2013
Development of innovative computer software to facilitate the setup and computation of water quality index
.
J. Environ. Health Sci. Eng.
11
(
1
),
1
.
http://dx.doi.org/10.1186/2052-336X-11-1
.
Praus
P.
2019
Principal component weighted index for wastewater quality monitoring
.
Water
11
(
11
).
http://dx.doi.org/10.3390/w11112376
.
Qin
G.
,
Liu
J.
,
Xu
S.
&
Sun
Y.
2021
Pollution source apportionment and water quality risk evaluation of a drinking water reservoir during flood seasons
.
Int. J. Environ. Res. Public Health
18
(
4
).
http://dx.doi.org/10.3390/ijerph18041873
.
Razmkhah
H.
,
Abrishamchi
A.
&
Torkian
A.
2010
Evaluation of spatial and temporal variation in water quality by pattern recognition techniques: a case study on Jajrood River (Tehran, Iran)
.
J. Environ. Manage.
91
(
4
),
852
860
.
http://dx.doi.org/10.1016/j.jenvman.2009.11.001
.
Roselli
L.
,
Fabbrocini
A.
,
Manzo
C.
&
D'Adamo
R.
2009
Hydrological heterogeneity, nutrient dynamics and water quality of a non-tidal lentic ecosystem (Lesina Lagoon, Italy)
.
Estuarine Coastal Shelf Sci.
84
(
4
),
539
552
.
http://dx.doi.org/10.1016/j.ecss.2009.07.023
.
Shen
Z.
,
Hong
Q.
,
Yu
H.
&
Liu
R.
2008
Parameter uncertainty analysis of the non-point source pollution in the Daning River watershed of the Three Gorges Reservoir Region, China
.
Sci. Total Environ.
405
(
1–3
),
195
205
.
http://dx.doi.org/10.1016/j.scitotenv.2008.06.009
.
Shen
Z.
,
Qiu
J.
,
Hong
Q.
&
Chen
L.
2014
Simulation of spatial and temporal distributions of non-point source pollution load in the Three Gorges Reservoir Region
.
Sci. Total Environ.
493
,
138
146
.
http://dx.doi.org/10.1016/j.scitotenv.2014.05.109
.
State Environmental Protection Administration
2002
Standard Methods for the Examination of Water and Wastewater(Version 4)
.
China Environmental Science Press
,
Beijing
,
China
.
Tian
L.
,
Li
H.
,
Li
F.
,
Li
X.
,
Du
X.
&
Ye
X.
2018
Identification of key influence factors and an empirical formula for spring snowmelt-runoff: a case study in mid-temperate zone of northeast China
.
Sci. Rep.
8
(
1
),
16950
.
http://dx.doi.org/10.1038/s41598-018-35282-x
.
Tong
X.
,
Zhou
Y.
,
Liu
J.
,
Qiu
P.
&
Shao
Y.
2021
Non-point source pollution loads estimation in Three Gorges Reservoir Area based on improved observation experiment and export coefficient model
.
Water Sci. Technol.
.
http://dx.doi.org/10.2166/wst.2021.508
.
Uddin
M. G.
,
Nash
S.
&
Olbert
A. I.
2021
A review of water quality index models and their use for assessing surface water quality
.
Ecol. Indic.
122
.
http://dx.doi.org/10.1016/j.ecolind.2020.107218
.
Venkatramanan
S.
,
Chung
S. Y.
,
Ramkumar
T.
,
Rajesh
R.
&
Gnanachandrasamy
G.
2015
Assessment of groundwater quality using GIS and CCME WQI techniques: a case study of Thiruthuraipoondi city in Cauvery deltaic region, Tamil Nadu, India
.
Desalin. Water Treat.
57
(
26
),
12058
12073
.
http://dx.doi.org/10.1080/19443994.2015.1048740
.
Wang
Z.
,
Wang
T.
,
Liu
X.
,
Hu
S.
,
Ma
L.
&
Sun
X.
2020
Water level decline in a reservoir: implications for water quality variation and pollution source identification
.
Int. J. Environ. Res. Public Health
17
(
7
).
http://dx.doi.org/10.3390/ijerph17072400
.
Wang, X., Li, Z. & Li, M. 2018 Impacts of climate change on stream flow and water quality in a drinking water source area, Northern China. Environmental Earth Sciences 77 (11). http://dx.doi.org/10.1007/s12665-018-7581-5
.
Wu
D.
,
Yan
H.
,
Shang
M.
,
Shan
K.
&
Wang
G.
2017a
Water eutrophication evaluation based on semi-supervised classification: a case study in Three Gorges Reservoir
.
Ecol. Indic.
81
,
362
372
.
http://dx.doi.org/10.1016/j.ecolind.2017.06.004
.
Wu
Z.
,
Zhang
D.
,
Cai
Y.
,
Wang
X.
,
Zhang
L.
&
Chen
Y.
2017b
Water quality assessment based on the water quality index method in Lake Poyang: the largest freshwater lake in China
.
Sci. Rep.
7
(
1
),
17999
.
http://dx.doi.org/10.1038/s41598-017-18285-y
.
Yan
M.
,
Chen
S.
,
Huang
T.
,
Li
B.
,
Li
N.
,
Liu
K.
,
Zong
R.
,
Miao
Y.
&
Huang
X.
2020
Community compositions of phytoplankton and eukaryotes during the mixing periods of a drinking water reservoir: dynamics and interactions
.
Int. J. Environ. Res. Public Health
17
(
4
).
http://dx.doi.org/10.3390/ijerph17041128
.
Ye
R.
,
Liu
L.
,
Wang
Q.
,
Ye
X.
,
Cao
W.
,
He
Q.
&
Cai
Y.
2017
Identification of coastal water quality by multivariate statistical techniques in two typical bays of northern Zhejiang Province, East China Sea
.
Acta Oceanolog. Sin.
36
(
2
),
1
10
.
http://dx.doi.org/10.1007/s13131-017-0981-7
.
You
Q.
,
Fang
N.
,
Liu
L.
,
Yang
W.
,
Zhang
L.
&
Wang
Y.
2019
Effects of land use, topography, climate and socio-economic factors on geographical variation pattern of inland surface water quality in China
.
PLoS One
14
(
6
),
e0217840
.
http://dx.doi.org/10.1371/journal.pone.0217840
.
Zeng
J.
,
Huang
G.
,
Luo
H.
,
Mai
Y.
&
Wu
H.
2019
First flush of non-point source pollution and hydrological effects of LID in a Guangzhou community
.
Sci. Rep.
9
(
1
),
13865
.
http://dx.doi.org/10.1038/s41598-019-50467-8
.
Zhang
T.
,
Ni
J.
&
Xie
D.
2016
Assessment of the relationship between rural non-point source pollution and economic development in the Three Gorges Reservoir Area
.
Environ. Sci. Pollut. Res. Int.
23
(
8
),
8125
8132
.
http://dx.doi.org/10.1007/s11356-016-6344-7
.
Zhao
P.
,
Li
Z.
,
Zhang
R.
,
Pan
J.
&
Liu
Y.
2020
Does water diversion project deteriorate the water quality of reservoir and downstream? A case-study in Danjiangkou reservoir
.
Global Ecol. Conserv.
24
.
http://dx.doi.org/10.1016/j.gecco.2020.e01235
.
Zotou
I.
,
Tsihrintzis
V. A.
&
Gikas
G. D.
2020
Water quality evaluation of a lacustrine water body in the Mediterranean based on different water quality index (WQI) methodologies
.
J. Environ. Sci. Health. Part A Toxic/Hazard. Subst. Environ. Eng.
55
(
5
),
537
548
.
http://dx.doi.org/10.1080/10934529.2019.1710956
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/).