Environmental Earth Sciences (2021) 80:368

https://doi.org/10.1007/s12665-021-09670-y

THEMATIC ISSUE

Surface water quality assessment based on the Integrated Water

Quality Index in the Maozhou River basin, Guangdong, China
Tao Lin1,2   · Huiqing Yu1,2 · Qi Wang1,2 · Lin Hu1,2 · Jing Yin1,2

Received: 20 January 2021 / Accepted: 22 April 2021

© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021

Abstract
Rivers are a vital component of both urban and rural aquatic ecosystems, and the rapidly increasing pollution in rivers
severely threatens the security of these ecosystems. In this study, the surface water quality in the Maozhou River basin,
Guangdong Province, China, was assessed from 2018 to 2020 using multivariate statistical techniques and the Integrated
Water Quality Index (IWQI). In addition, spatial trends in surface water quality were studied using a geographic information
system. The results indicated that the water quality in 82.17% of the studied section met the Class V standard for surface
water quality, with IWQI values ranging from 12.157 to 3.650. The surface water quality was clustered into eight groups
and further divided into unsuitable (low quality) and suitable (acceptable quality) according to standard surface water qual-
ity thresholds (China in Environmental quality standards for surface water, 2002). Four groups (G1, G2, G3, and G4) were
classified as unsuitable, because they fell short of the Class V standard for surface water quality. In these groups, industrial
sewage, endogenous pollution, domestic sewage, and rainfall runoff, were the primary sources of pollution, and the main
background pollutants for the water quality target of the functional zones were COD, N ­ H3-N, TP, and LAS. The other four
groups were classified as suitable. In these groups, endogenous pollution and rainfall runoff were the primary sources of pol-
lution, and the main background pollutants for the water quality target of the functional zones were ­NH3-N and TP. Among
them, ­NH3-N and LAS were recognized as responsive and sensitive to the surface water quality and spatio-temporal vari-
ability. Owing to pollution treatment and management measures undertaken by the Chinese government, the black-odorous
water in the Maozhou River basin has disappeared, and the water quality in the Maozhou River basin has been maintained
at the “medium and good” level. However, the surface water quality in the estuary region and the southwest tributary in the
basin requires further improvement. This calls for further efforts to improve surface water quality and to properly deal with
various sources of pollution in the watershed. This combined method has proved to be effective for surface water quality
evaluation and management at river or basin scales. The results of this work are expected to provide a scientific foundation
for aquatic ecosystem management and planning.

Keywords  Integrated Water Quality Index (IWQI) · Multivariate statistical techniques · Geographic information system
(GIS) · Black-odorous water · Maozhou River basin

Introduction
This article is part of a Topical Collection in Environmental Earth
Sciences on Groundwater quality and contamination and the Adequate and high-quality water is vital for the sustainable
application of GIS, guest edited by Narsimha Adimalla and Hui development of human society and ecosystems (Vorosmarty
Qian. et al. 2010). Increases in industrialization, expansion of
urbanization, and rapid economic development have caused
* Tao Lin
[email protected] severe pollution of surface water bodies (Madan et al. 2020).
River water, which is the lifeblood of urban ecosystems,
1
CCCC First Highway Consultants Co., Ltd, No. 63 is under enormous pressure to fulfill the demands of soci-
Technology Second Road, Xi’an 710075, Shaanxi, China ety, especially in economically developing regions, such
2
Xi’an Zhongjiao Environmental Engineering Co., Ltd, as the Guangdong–Hong Kong–Macao Greater Bay Area
No. 205 Technology Four Road, Xi’an 710065, Shaanxi, (GBA). The rapid economic development of the GBA has
China

13
Vol.:(0123456789)
 368   Page 2 of 16 Environmental Earth Sciences (2021) 80:368

engendered environmental challenges, which constrain the analysis (HCA), are effective tools for water quality data
sustainable development of the region and thus affect its goal analysis. These methods have been proved useful in the
to construct a first-class international bay area (Cai et al. evaluation of hydrogeochemical processes and the effec-
2020; Weng et al. 2020). Thus, there is an urgent need to tive identification of pollution sources (Ana et al. 2019).
formulate effective evaluation and management strategies PCA converts large datasets of correlated variables into
for sustainable utilization and protection of the environment, a small number of uncorrelated factors (Howladar et al.
particularly surface water resources. Previous studies in the 2017). HCA employs algorithms to divide similar objects
area have focused on heavy metal/metalloid stabilization into groups (Islam et al. 2017). The Integrated Water Quality
and shallow groundwater evolution. However, surface water Index (IWQI) can be used to qualitatively and quantitatively
quality classification and spatial quality characterization, evaluate water quality (Xu 2005). A geographic information
with the purpose of identifying the most vulnerable regions system (GIS) can provide a clear overview of the overall
and their pollution sources, has not been addressed (Fang water quality in an area based on the important water quality
et al. 2018; Shi et al. 2018; Liu et al. 2019; Chen et al. 2020). parameters (Abbasi and Abbasi 2012).
To ensure economic growth and social development, This study comprehensively evaluates surface water qual-
it is essential to maintain surface water quality (Gao ity in the target area using GIS-based multivariate statisti-
et al. 2020; Nong et al. 2020). Increases in urbanization, cal techniques and the IWQI. In addition, to evaluate the
industrialization, and agricultural activities have severely changes in hydrological conditions in space and time, all
impaired the quality of both surface water and ground- parameters of surface water data are analyzed using the Sta-
water in many countries and regions (Adimalla and Qian tistical Package for the Social Sciences (SPSS) program.
2019; Zhang et al. 2020a, b, c). Therefore, it is important Water quality evaluation results corresponding to sampling
to implement effective management measures to prevent points are obtained using the IWQI, and the spatial charac-
and control the deterioration of water quality (Peyman teristics of surface water quality are mapped using a GIS.
et al. 2017). Surface water quality assessment is a process This combined method effectively achieves water quality
including, but not limited to, both qualitative studies and evaluation at a large scale.
quantitative evaluations, which include classification of
water quality and determination of pollution types (Wu
et al. 2004; Lin et al. 2017; Edegbene et al. 2019). Surface
water quality evaluation based on the field sampling and Site description
chemical methods is essential to understand water quality
and control water pollution (Sun et al. 2016; Sany et al. Study area
2019). During the past few decades, methods involving
the water quality index (WQI) and fuzzy comprehensive The Maozhou River basin is located in the northwest part
assessment have been widely used in water quality assess- of Shenzhen, Guangdong Province, China. The length of
ment (Ramos et al. 2016; Hamlat and Guidoum 2018; Wu the main stream is 31.29 km and the drainage area of the
et al. 2019). The quantitative estimation of water qual- basin is 388 ­km2 (Fig. 1). There are 41 rivers in the basin,
ity parameters via multispectral remote sensing analysis including 1 main stream, 23 first-order tributaries, and 17 s-
has yielded relatively accurate results, but this approach or third-order tributaries. The Maozhou River, which is the
is time-consuming and expensive (Harvey et al. 2015). main stream, flows gently and is affected by high tides. The
Although numerous researchers have focused on the appli- annual average precipitation in the Maozhou River basin
cability of surface water quality assessment methods, they is 1938.6 mm; it gradually decreases from west to east.
did not map the spatial variation of surface water quality Approximately 80% of the annual precipitation is received
(e.g., Babiker et al. 2007; Eilander et al. 2014; Naveedul- from April to September. This basin has a subtropical oce-
lah et al. 2016; Nong et al. 2020). Although previous stud- anic monsoon climate, with an annual average temperature
ies have independently evaluated surface water quality of 22 °C (Zhou et al. 2010; Shi et al. 2020). Industrial waste
using various assessment methods under different spatio- and sewage from a large number of factories (e.g., electro-
temporal conditions, the spatio-temporal variations in the plating, printed circuit board manufacturing, optoelectronic
surface water quality index have seldom been considered device manufacturing) and residential areas distributed
(Lin et al. 2017). It is challenging to analyze hydrogeo- along the river are discharged into the water, causing the
chemical datasets of surface water at large spatial and tem- deterioration of the aquatic ecosystem of the Maozhou River
poral scales, and the use of traditional evaluation methods (Liu et al. 2019; Zhang et al. 2020a, b, c). Moreover, the
may lead to unnecessary and redundant calculations and estuary of the river is low-lying and tidal crests are frequent.
evaluations. Multivariate statistical methods, such as prin- As a result, the river ecology was irreversibly destroyed and
cipal component analysis (PCA) and hierarchical cluster the water became black and odorous.

13
Environmental Earth Sciences (2021) 80:368 Page 3 of 16  368

Fig. 1  Location map of water samples in part of Maozhou River Basin, China

13
 368   Page 4 of 16 Environmental Earth Sciences (2021) 80:368

Sampling and data collection original variables to the greatest extent; thus, explaining
the essence of the variables (Howladar et al. 2017). HCA is
Surface water samples were collected from the river channel widely used in cluster grouping; in this study, it was adopted
monitoring sections in the Maozhou River basin at weekly to obtain a comprehensive understanding of the water qual-
intervals from 2018 to 2020. The monthly average water ity via the discretization of the samples into different groups
quality information was obtained from the weekly water (Lambraski et al. 2004). In essence, the process of cluster
quality data in the study area, which is based on the basic analysis involves the calculation of the distance between
parameters of the Environmental Quality Standards for Sur- each hydrochemical parameter of the surface water sam-
face Water (China 2002) and the protocols outlined in the ples. The greater the similarity in water quality, the greater
Guidance on Sampling Techniques (China 2009). Figure 1 the likelihood that the samples will be in the same water
shows the locations of surface water sampling points in the quality cluster (Shyu et al. 2011; Chin et al. 2012). HCA
study area. Maozhou River is the largest river in Shenzhen helps to maximize the similarity and minimize the difference
city and adjacent to an estuary with complex hydrodynamic between samples within a group of water samples (Varol
conditions (Chen et al. 2019; Liu et al. 2019). Nutrient and et al. 2012).
organic pollutants are ubiquitous pollutants, and there is no
heavy metal factory in the area (Shi et al. 2018). In this
study, the main indexes influencing river water quality were Surface water quality index
selected as evaluation parameters, based on the relevant lit-
erature and research results (Jiang et al. 2019; Zhang et al. The IWQI consists of an integer and three or four decimal
2020a, b, c). In all, nine water quality parameters were places (Xu 2005; Lin et al. 2017). It is calculated using the
selected, namely, flow velocity (m/s), water depth (cm), dis- following formula:
solved oxygen (DO, mg/L), chemical oxygen demand (COD,
IWQI = X1 ⋅ X2X3X4, (1)
mg/L), ammonia nitrogen (­ NH3-N, mg/L), total phosphorus
(TP, mg/L), fluoride (­ F−, mg/L), oxidation–reduction poten- n
tial (ORP, mV) and linear alkylbenzene sulfonates (LAS, 1∑
X1 ⋅ X2 = Pi, (2)
mg/L), which are anionic surfactants. During sample col- n i=1
lection, flow velocity, water depth, and DO were determined
using a portable flow velocity meter (LS1206B), portable where X1 is the water quality category; X2 is the position of
ultrasonic sounder (ZMSS-100), and YSI multiparam- the comprehensive water quality within the range of class
eter probes (556MPS), respectively. Samples to assess the X1 ; X3 is the number of single indices that are inferior to the
remaining parameters were collected in polyethylene bottles target of the water environment functional zone. If X3 = 1, it
and stored in refrigerators. All sample bottles were leak- means that one of the indexes included in the comprehensive
proof and were labeled with detailed information to prevent water quality assessment falls short of the functional zone
misdiagnosis. Geographic coordinates were also noted for target; X4 is the result of the comparison between the com-
each sampling point using the global positioning system prehensive water quality category and the functional zone
(GPS). All water samples were immediately transported to category. If X4 = 1, it indicates that the surface water quality
the laboratory, and analyses were conducted within 24 h of is inferior to the water quality target of the functional zones;
collection. Pi indicates the parameter value of a single water quality
index; and n is the number of parameters.
The single water quality index was computed using the
Methodology following methods. Only dissolved oxygen is a decreasing
index, and was calculated using Eq. 3. The other water qual-
Data processing ity indicators, which are incremental indicators, were calcu-
lated using Eq. 4.
Detailed surface water quality data were analyzed by con- 𝜌DO,k+1 −𝜌DO,i
ducting analysis of variance (ANOVA), PCA, and HCA for PDO = KDO +
𝜌DO,k+1 − 𝜌DO,k
, (3)
all parameters using SPSS (Dassi 2011; Zhang et al. 2014;
Xu et al. 2019a, b, c, 2021). The influence of two independ-
𝜌i −𝜌i,k
ent factors (i.e., time and space) on the surface water quality Pi = Ki + . (4)
was evaluated using one-way ANOVA (Varol 2019). PCA is 𝜌i,k+1 − 𝜌i,k
often used to extract a few potential common factors from
When the surface water quality parameter exceeded
numerous observable variables (Xu et al. 2019c; Zhang et al.
the threshold of the Class V standard (Table  2), the
2020a, b, c). These factors explain the information of the

13
Environmental Earth Sciences (2021) 80:368 Page 5 of 16  368

incremental water quality index and dissolved oxygen GIS‑IWQI method

index were calculated using Eqs. 5 and 6.
𝜌DO,k5 −𝜌DO,i The IQWI value of the Maozhou River basin was employed
PDO = 6 + , (5) to draw a water quality map, following the procedure pro-
𝜌DO,k5
posed by Babiker et al. (2007). The entire process is carried
out in three main steps viz., the date of surface water sample
𝜌i −𝜌i,5 analysis, surface water quality evaluation, and GIS-based
Pi = 6 + (6)
𝜌i,5 surface water quality mapping.

where Ki represents the ith water quality category, accord-

ing to environmental quality standards for surface water; its Results and discussion
value ranges from 1 to 6. ρi is the measured value of the ith
water quality index. ρ (i, k) is the standard value of the Ki class Surface water quality characteristics
water quality category.
In this study, based on the IWQI value (Table 2), the Table 2 indicates that the annual average water depth was
surface water quality was classified into seven grades, greater than 32.49 cm, and the values gradually declined
namely, excellent (1–2), very good (2–3), good (3–4), during monitoring period. The annual average flow rate
medium (4–5), acceptable (5–6), inferior to Class V but exceeded 0.08 m/s; the maximum value was recorded in
not black and odorous (6–7), and inferior to Class V and 2018 and the minimum value was recorded in 2020. This
black and odorous (> 7). This classification is in accord- shows that the hydrodynamic conditions in the basin gradu-
ance with China’s actual surface water quality manage- ally deteriorated during the monitoring period.
ment standards (Table 1). The annual average DO concentrations for each year
exceeded the thresholds of the Class V standard (2 mg/L).
The lowest DO concentration (0.39 mg/L) was observed in
2018; the DO concentrations exceeded 2 mg/L in 98.6% of
the sections in 2019 and in all sections in 2020. The DO con-
Table 1  The IWQI range and status of surface water quality
centrations in the channel demonstrated a significant increas-
IWQI Rank Water quality status ing trend. The annual average COD concentration decreased
from 2018 to 2020; the highest average COD concentration
1 < X1·X2 ≤ 2 Class I Excellent
of 54.26 mg/L was observed in 2018, and the annual aver-
2 < X1·X2 ≤ 3 Class II Very good
age COD concentration was lower than the thresholds of
3 < X1·X2 ≤ 4 Class III Good
the Class V standard (40 mg/L) from 2019 to 2020. The
4 < X1·X2 ≤ 5 Class IV Medium
COD concentration is an indicator of the level of pollution of
5 < X1·X2 ≤ 6 Class V Acceptable
water by organic matter. The decreasing trend indicated that
6 < X1·X2 ≤ 7 Inferior class V Inferior class V but not black stink
the degree of organic pollution was reduced in the Maozhou
X1·X2 > 7 Inferior class V Inferior class V and black stink
River basin.

Table 2  Comparison of the variations of the water quality parameters from 2018 to 2020
Parameters Thresholds of the 2018 2019 2020
Class V ­standardsa
Avg. ± SD Min Max Avg. ± SD Min Max Avg. ± SD Min Max

Flow (m/s) N/A 0.15 ± 0.09 0.02 0.51 0.09 ± 0.06 0 0.28 0.08 ± 0.06 0 0.23
Depth (cm) N/A 65.62 ± 63.11 8.42 300 35.51 ± 13.09 6.71 57.67 32.49 ± 14.7 6.36 84.45
DO (mg/L)  ≥ 2 mg/L 2.74 ± 1.75 0.39 7.12 4.21 ± 0.99 1.74 5.92 5.08 ± 0.73 3.23 9.27
COD (mg/L)  ≤ 40 mg/L 54.26 ± 39.61 17.99 205.4 31.25 ± 10.46 18.81 73.12 25.18 ± 2.18 21.56 37.11
NH3-N (mg/L)  ≤ 2 mg/L 13.78 ± 10.88 0.99 45.8 5.63 ± 4.07 0.57 18.94 1.66 ± 0.84 0.57 4.79
TP (mg/L)  ≤ 0.4 mg/L 2.11 ± 3.48 0.16 29.04 0.76 ± 0.57 0.12 3.15 0.27 ± 0.26 0.07 2.13
F− (mg/L)  ≤ 1.5 mg/L 1.20 ± 3.55 0.24 32.32 0.59 ± 0.38 0.30 2.94 0.57 ± 0.18 0.3 1.19
LAS (mg/L)  ≤ 0.3 mg/L 0.65 ± 0.53 0.08 2.31 0.26 ± 0.17 0.10 1.04 0.13 ± 0.05 0.09 0.37
ORP (mv) N/A 245.10 ± 47.92 117.93 335.42 286.40 ± 18.54 218.92 328.92 300.89 ± 14.27 252.34 326.59

Avg. average, SD standard deviation

a
 Standards from the Environmental Quality Standards for Surface Water (China 2002)

13
 368   Page 6 of 16 Environmental Earth Sciences (2021) 80:368

The annual average ­NH3-N concentration was 1.66 mg/L dissolved oxygen decreases with increasing depth. Decom-
in 2020, which was lower than the corresponding values position of organic matter and nutrients consumes dissolved
in 2018 and 2019. The temporal evolution of the ­NH3-N oxygen, creating an anaerobic environment and causing the
concentration exhibited a decreasing trend with fluctua- water quality to deteriorate. Lv et al. (2014) showed that
tions from 2018 to 2020. The N ­ H3-N concentration detected black-odorous water can be formed at depths of 0.9–1.5 m.
every year exceeded the threshold for the Class V standard The hydrodynamic parameters showed moderate or weak
(2 mg/L), indicating that the ­NH3-N concentrations could correlations with the chemical water quality parameters,
cause the surface water quality to be classified as inferior to indicating that they exerted an influence on surface water
Class V in some sections. The annual average and maximum quality, but were not the main influencing factors. DO had a
TP concentrations decreased from 2018 to 2020, although strong and moderately negative correlation relationship with
the maximum and average TP concentrations exceeded the nutrients and organic pollutants, respectively, indicating that
threshold for the Class V standard (0.4 mg/L) from 2018 DO was the main factor influencing surface water quality.
to 2019. The TP parameter exhibited a strong tendency to
cause the water quality to be classified as inferior to Class V. Principal component analysis (PCA)
­NH3-N and TP can be used as parameters to characterize the
level of water pollution by nutrients (Mao et al. 2019). This PCA is a multivariate statistical method that can be
indicated that the surface water environment was polluted used to identify the correlation between variables (Xiao
by nutrients in the study area. et  al. 2019). The principal components (PCs) among
The annual average concentrations of ­F− and LAS showed the hydrochemical characteristics of the surface water
minimal variations, ranging from 0.57 to 1.2 mg/L and 0.13 samples were identified using a scree plot (Fig. 3) and
to 0.65 mg/L, respectively. The maximum LAS concentra- PC loadings (Table 3) (Zhang et al. 2019). Factors with
tion exceeded the threshold of the Class V standard for sur- eigenvalues greater than 1 were selected as PCs (Xu et al.
face water quality (0.3 mg/L), indicating that the surface 2019c). PCA generates three components, namely, PC1,
water quality and the survival of aquatic organisms could be PC2, and PC3; each PC loading can be used to explain
adversely affected by organic pollutants. The annual average the surface water quality characteristics (Xu et al. 2019b).
value of ORP increased year by year, indicating an increase PC1, with the maximum variance (48.8%), dominated the
in the tendency of pollutants to undergo oxidation in the hydrochemical characteristics of the surface water sam-
water. Thus, nitrogen, phosphorus, and other nutrients in ples. PC1 had positive loadings on COD (0.912), ­NH3-N
water were more likely to be oxidized into inorganic sub- (0.957), LAS (0.9), and TP (0.843), and a negative load-
stances, reducing the risk of water eutrophication. ing on DO (− 0.792). The high loading of PC1 on N ­ H3-N
and COD indicated that organic pollution and nutrient
Correlation analysis of water quality index concentrations impact surface water quality. Nutrient and
organic pollutants mainly originate from domestic sew-
The relationship between different water parameters is rep- age and rainfall runoff in the study area (Fragaa et al.
resented by Pearson’s correlation coefficient (r). The cor- 2016; Liang et al. 2020). In particular, large amounts of
relation is considered strong if the r value is greater than domestic sewage are discharged from residential areas to
0.7, whereas it is considered weak if the r value is lower the river through municipal pipelines, and factories along
than 0.3. A moderate correlation is considered to exist when both banks of the river dump industrial garbage and sew-
r value lies between 0.3 and 0.7 (Emenike et al. 2018; Xu age into the river. PC2 accounted for 20.9% of the total
et al. 2019b; Zhang et al. 2020a, b, c). Figure 2 demonstrates variance and was dominated by ­F − (0.743), indicating
that DO has a strong relationship with ­NH3-N, TP, and ORP. that water pollution in the basin was affected by indus-
Similarly, ­NH3-N has a strong positive relationship with trial wastewater (Liu et al. 2019). There are more than
COD and TP. This indicates that the occurrence of water 12,000 enterprises in the Maozhou River basin (Baoan
hypoxia is the main factor influencing the increase in organic Area), among which 274 are key polluting enterprises.
pollutant (i.e., COD, LAS) and nutrient (i.e., N ­ H3-N, TP) Industrial land accounts for 43% of the total land area,
concentrations in river water. Under the strongly reducing and industrial water consumption accounts for 61% of
environment, the sediment releases large amounts of nutri- the total water consumption. In the industrial area, the
ents and organic pollutants to the overlying water, which large number of polluting industrial activities such as
causes the river water quality to progressively deteriorate electroplating, printed circuit board manufacturing, and
(McGregor and Fabbro 2000; Li et al. 2021). Water depth optoelectronic device manufacturing, and industrial pro-
exhibits a moderate correlation with flow velocity, DO, and duction has greatly contributed to the pollution in the
ORP, indicating that the water depth has a direct impact on Maozhou River basin (Lou et al. 2020). PC3 accounted
the dissolved oxygen concentration. The concentration of for 13.4% of the total variance and had a high loading on

13
Environmental Earth Sciences (2021) 80:368 Page 7 of 16  368

Fig. 2  Relationships between different indices in this study region

13
 368   Page 8 of 16 Environmental Earth Sciences (2021) 80:368

Surface Water (China 2002), and was not analyzed in this

study. The remaining six variables were selected as the
important indexes of surface water quality evaluation in
this work.

Analysis of variance (ANOVA)

The data obtained from the surface water samples collected

from the Maozhou River trunk and branch streams were
analyzed using one-way ANOVA (Tables 4 and 5). During
the monitoring period, COD and LAS exhibited a statisti-
cally significant change trend (p < 0.01), whereas the other
parameters did not. The surface water quality parameters
exhibited statistically significant change trends that varied
in space, i.e., certain sections of the drainage basin exhibited
statistically significant change trends. The p value can be
used to determine which parameters constitute independent
factors (Grunsky et al. 2009). The one-way ANOVA results
Fig. 3  Scree plot from PCA loadings scores for dataset of water sam- indicated that ­NH3-N and LAS were responsive and sensitive
ples towards the surface water quality and spatio-temporal vari-
ability. The surface water quality parameters demonstrated
Table 3  The water quality rotation factor load matrix spatio-temporal variations, and surface water samples could
Evaluation index Principal components
be grouped according to the differences in their hydrochemi-
cal characteristics in this study.
1 2 3

Flow − 0.369 0.07 − 0.083 Hierarchical cluster analysis (HCA)

Depth 0.087 − 0.461 0.841
DO − 0.792 0.311 − 0.171 HCA was used to cluster water samples with similar hydro-
COD 0.912 0.023 − 0.151 chemical characteristics into groups. A visual overview of
NH3-N 0.957 0.008 − 0.066 the clustering process is presented as a dendrogram. Bore-
TP 0.843 0.356 0.075 holes that exhibit similarities are primarily categorized into
F− 0.119 0.743 0.442 three distinguishable groups. According to the dendrogram,
LAS 0.9 − 0.115 − 0.134 water samples were mainly categorized into eight groups by
ORP − 0.109 0.498 0.148 SPSS20.0 based on the similarity in this study (Table 6). The
average values of the hydrochemical parameters of all sam-
ples in a group are used to represent the entire group. On this
depth (0.84), suggesting that the deep water area is also basis, the average surface water quality value after grouping
one of the factors influencing the deterioration of water is obtained (Table 7). The average value of the parameters in
quality (Lv et al. 2014). The absolute values of factor each group was evaluated in terms of the IWQI. The evalua-
loading for COD, N ­ H 3-N, TP, ­F −, LAS, DO, and depth tion results of each group were assigned to their correspond-
were greater than 0.7 in the rotational component matrix ing sample points (Table 6).
table. The depth is not clearly classified as a water qual- Group 1 was composed of surface water samples S9, S16,
ity standard in the Environmental Quality Standards for S17, S18, and S20, which were obtained from tributaries in

Table 4  ANOVA of spatial Time Sig.(p)

variation among section
DO COD NH3-N TP F− LAS

2018 0.002** 0.574 0.002** 0.081 0.153 0.000**

2019 0.460 0.980 0.044* 0.014* 0.101 0.001**
2020 0.396 0.476 0.002** 0.212 0.002** 0.021*

*Significance value p < 0.05

**Significance level p < 0.01

13
Environmental Earth Sciences (2021) 80:368 Page 9 of 16  368

Table 5  ANOVA of spatial Section Significance

variation among 2018–2020
DO COD NH3-N TP F− LAS

S1 0.988 0.160 0.182 0.517 0.000** 0.623

S2 0.140 0.085 0.180 0.000** 0.324 0.012*
S3 0.055 0.462 0.257 0.128 0.132 0.032*
S4 0.288 0.009** 0.094 0.029 0.000** 0.003**
S5 0.206 0.005** 0.140 0.005** 0.000** 0.012*
S6 0.813 0.091 0.721 0.330 0.070 0.001**
S7 0.468 0.325 0.077 0.039* 0.008** 0.001**
S8 0.216 0.000** 0.121 0.024* 0.027* 0.134
S9 0.934 0.568 0.735 0.718 0.000** 0.349
S10 0.085 0.051 0.131 0.048* 0.246 0.004**
S11 0.115 0.008** 0.171 0.153 0.000** 0.005**
S12 0.022* 0.074 0.003** 0.007** 0.223 0.000**
S13 0.125 0.022* 0.001** 0.003** 0.013* 0.023*
S14 0.143 0.021* 0.001** 0.023* 0.293 0.325
S15 0.275 0.003** 0.004** 0.003** 0.000** 0.124
S16 0.480 0.025* 0.629 0.148 0.000** 0.909
S17 0.174 0.002** 0.007** 0.002** 0.552 0.014*
S18 0.373 0.026* 0.234 0.027* 0.017* 0.355
S19 0.523 0.000** 0.000** 0.000** 0.017* 0.000**
S20 0.975 0.469 0.723 0.695 0.437 0.317
S21 0.298 0.000** 0.013* 0.025* 0.736 0.001**
S22 0.465 0.016* 0.259 0.431 0.029* 0.012*
S23 0.984 0.278 0.452 0.001** 0.015* 0.012*
S24 0.950 0.030* 0.185 0.513 0.098 0.001**

*Significance value p < 0.05

**Significance level p < 0.01

the upper and middle reaches of the Maozhou River during that exhibited the minimum concentrations of ­NH3-N, TP,
the nonflood season in 2018. This group exhibited the high- ­F−, and LAS in different seasons in 2020, which were below
est concentrations of COD, N ­ H3-N, TP, F­ −, and LAS, and the thresholds of the Class V standard. Group 7 had the
the lowest concentrations of DO, which indicates that this largest number of samples; these samples were obtained
group had the most serious water pollution in the study area. from different sections and during various periods. Group 8
Group 2 was composed of surface water samples S9, S11, clustered water samples that exhibited the highest concen-
S16, S17, and S18, which were obtained from tributaries trations of DO and the lowest concentrations of COD from
in the middle reaches of the Maozhou River in 2018. The the study area. Generally, Group 1 samples were character-
average concentrations of COD, ­NH3-N, TP, and LAS in this ized by high COD, ­NH3-N, TP, and LAS concentrations and
group were relatively high, but the average concentrations of low DO concentrations, indicating that these sampling sites
DO and LAS exceeded the thresholds of the Class V stand- were considerably affected by domestic sewage and non-
ard. The degree of water pollution was second only to the point source pollution. Group 3 samples were characterized
that of Group 1. Group 3 clustered samples from Maozhou by high ­F− concentrations, indicating that these sampling
River tributaries in which the maximum concentrations of sites were greatly affected by industrial sewage. However,
­F− were most serious. Group 4 mainly clustered water qual- the average concentrations of COD, N ­ H3-N, TP, and LAS
ity samples from the trunk (S1 and S7) and tributaries (S8, also exceeded the thresholds for the Class V standard (China
S14, S17, S23, and S24) in 2018. Group 5 included water 2002), indicating that the river water quality was also threat-
samples obtained from parts of the trunk and tributaries dur- ened by domestic sewage and nonpoint source pollution.
ing 2018–2020. In these samples, the average concentrations Group 6 samples were characterized by low N ­ H3-N, TP,
of DO, COD, and F ­ − were below the thresholds of the Class ­F−, and LAS concentrations, indicating that the stoppage
V standard, whereas the concentrations of ­NH3-N, TP, and of domestic and industrial sewage had a significant effect
LAS exceeded these thresholds. Group 6 clustered samples and the river water quality had improved. Group 8 samples

13
 368   Page 10 of 16 Environmental Earth Sciences (2021) 80:368

Table 6  The corresponding samples of 8 groups

No Number Water samples

G1 17 S9-18-1~7, S9-18-9, S16-18-1, S17-18-1~5, S18-18-1, S20-18-1~2

G2 38 S9-18-8, S9-18-10~12, S10-18-1, S11-18-1~4, S16-18-2~7, S17-18-7, S17-18-10, S18-18-2, S18-18-10, S20-18-3~5, S20-18-7,
S20-18-10~12, S21-18-4, S22-18-1~2, S22-18-4, S22-18-6, S22-18-8, S9-19-1~3, S20-19-6~7
G3 49 S1-18-1, S1-18-7,S6-18-1~4, S8-18-12, S11-18-5, S11-18-7~8, S11-18-10~12, S15-18-8, S16-18-8, S16-18-10~11, S17-18-
8~9, S18-18-3~9, S20-18-6, S20-18-8, S21-18-6, S21-18-7, S22-18-3, S22-18-11, S23-18-2, S23-18-5, S24-18-4~8, S1-19-5,
S8-19-4, S11-19-1~3, S14-19-2, S16-19-1~5, S18-19-3, S20-19-1~3, S20-19-9, S21-19-9, S22-19-5~6, S24-19-1
G4 42 S1-18-2~45, S1-18-8~9, S6-18-12, S7-18-1, S7-18-12, S8-18-1, S8-18-3~5, S11-18-6, S14-18-1~4, S17-18-6, S17-18-11~12,
S18-18-12, S20-18-9, S21-18-1, S21-18-3~4, S21-18-8, S22-18-7, S23-18-1, S23-18-3~4, S23-18-6~9, S23-18-11~12, S24-
18-1, S24-18-9~12
G5 53 S1-19-1, S1-19-3~4, S1-19-6, S1-19-7, S4-19-3, S5-19-2, S6-19-1~2, S7-19-2, S8-19-1~3, S16-19-6~7, S16-19-9~12, S17-19-
1~4, S18-19-1~2, S18-19-5~6, S19-19-10, S20-19-4~5, S20-19-8, S20-19-11~12, S21-19-5~6, S21-19-10~12, S22-19-3~4,
S22-19-7~9, S22-19-12, S23-19-1~8, S24-19-2
G6 110 S1-20-1, S1-20-6~8, S1-20-10, S2-20-2, S2-20-5~6, S2-20-10, S3-20-1~2, S3-20-4, S3-20-6~7, S3-20-12, S4-20-1~2, S4-20-
4~5, S4-20-7, S4-20-10, S5-20-1, S5-20-3~5, S5-20-10, S6-20-4, S7-20-1, S7-20-3, S7-20-7~9, S7-20-11, S8-20-1, S8-20-
5~7, S8-20-9~11, S9-20-6, S9-20-10, S10-20-5~10, S11-20-5~7, S11-20-10, S12-20-1, S12-20-3, S12-20-5~7, S12-20-10,
S13-20-1~3, S13-20-9, S14-20-7~8, S14-20-10, S15-20-3~5, S16-20-3, S16-20-6, S16-20-8, S17-20-1~3, S18-20-1~2, S18-
20-4, S19-20-1, S19-20-4, S19-20-6~9, S20-20-1~2, S20-20-4~7, S21-20-1~4, S21-20-8~10, S22-20-1, S22-20-3, S22-20-6,
S22-20-8~9, S23-20-3, S23-20-7, S23-20-9~10, S24-20-1~3, S24-20-6, S24-20-10
G7 91 S2-18-1~3, S2-18-5~11, S3-18-1~9, S3-18-11~12, S4-18-1~3,S4-18-8, S5-18-2, S8-18-10~11, S10-18-9~11, S12-18-2, S12-18-
4, S12-18-7, S12-18-9~10, S13-18-1~4, S13-18-7, S13-18-10~12, S14-18-5~6, S14-18-10, S15-18-9~10, S19-18-1~2, S19-18-
12, S21-18-9, S22-18-12, S1-19-2, S2-19-2, S2-19-8, S3-19-2~3, S8-19-12, S9-19-5~6, S10-19-1~2, S12-19-3~4, S17-19-
6~10, S18-19-9, S18-19-12, S19-19-2, S1-20-2, S2-20-11, S4-20-11~12, S8-20-3, S9-20-2, S9-20-11, S10-20-11, S11-20-11,
S12-20-11, S13-20-6, S13-20-11, S14-20-1~3, S15-20-11, S16-20-11, S18-20-11, S19-20-1
G8 466 S1-18-10~12, S2-18-4, S2-18-12, S4-18-1, S4-18-3~12, S5-18-4~7, S5-18-9~12, S6-18-5~11, S7-18-2~5, S7-18-6~11, S8-18-2,
S8-18-6~9, S10-18-2~8, S10-18-12, S11-18-9, S12-18-1, S12-18-3, S12-18-5~6, S12-18-8, S12-18-11~12, S13-18-3, S13-
18-5~6, S13-18-8~9, S14-18-7~9, S14-18-11~12, S15-18-1~7, S15-18-11~12, S16-18-9, S16-18-12, S19-18-3~11, S21-18-5,
S21-18-10~12, S22-18-9~10, S23-18-10, S24-18-2~3, S1-19-8~12, S2-19-1~7, S2-19-9~12, S3-19-1, S3-19-4~12, S4-19-
1~2, S4-19-4~12, S5-19-1, S5-19-3~12, S6-19-3~12, S7-19-1, S7-19-3~12, S8-19-5~11, S9-19-4, S9-19-7~12, S10-19-3~12,
S11-19-4~12, S12-19-1~2, S12-19-5~12, S13-19-1~12, S14-19-1, S14-19-3~12, S15-19-1~12, S16-19-8, S17-19-5, S17-19-9,
S17-19-11~12, S18-19-4, S18-19-7~11, S19-19-1, S19-19-3, S19-19-5~12, S20-19-10, S21-19-1~4, S21-19-7, S21-19-8,
S22-19-1~2, S22-19-10~11, S23-19-9~12, S24-19-3~12, S1-20-3~5, S1-20-9, S1-20-11~12, S2-20-1, S2-20-3~4, S2-20-
7~9, S3-20-5, S3-20-8~11, S4-20-3, S4-20-6, S4-20-8, S4-20-9, S5-20-2, S5-20-6~9, S5-20-11~12, S6-20-1~3, S6-20-5~12,
S7-20-2, S7-20-4~6, S7-20-12, S8-20-2, S8-20-4, S8-20-8, S8-20-12, S9-20-1, S9-20-3~5, S9-20-7~9, S9-20-12, S10-20-1~4,
S10-20-12, S11-20-1~4, S11-20-8~9, S11-20-12, S12-20-12, S12-20-2, S12-20-4, S12-20-8~9, S13-20-4~5, S13-20-7~8, S13-
20-12, S14-20-4~6, S14-20-11~12, S15-20-1~2, S15-20-6~10, S15-20-12, S16-20-1~2, S16-20-4, S16-20-7, S16-20-9~10,
S16-20-12, S17-20-4~12, S18-20-3~5, S18-20-7~8, S18-20-10, S18-20-12, S19-20-2~3, S19-20-5, S19-20-10, S19-20-12,
S20-20-3, S20-20-8~9, S20-20-11~12, S21-20-5~7, S21-20-11~12, S22-20-2, S22-20-4, S22-20-7, S22-20-10~12, S23-20-
1~2, S23-20-4, S23-20-6, S23-20-8, S23-20-12, S24-20-4~5, S24-20-7~9, S24-20-11~12

Sxi-yj-n represented the samples of the xi section in the year yj and month n

Table 7  The mean value of No DO COD NH3-N TP F− LAS

each water quality indicator of 8
groups (unit: mg/L) G1 1.02 169.95 38.69 5.63 0.83 1.76
G2 2.10 99.98 26.24 3.20 0.80 1.41
G3 2.42 64.85 16.08 2.00 0.97 0.84
G4 2.74 43.86 13.23 1.59 0.79 0.59
G5 3.01 36.47 8.24 1.32 0.84 0.36
G6 5.08 27.05 1.59 0.26 0.58 0.13
G7 4.51 24.08 2.87 0.49 0.63 0.17
G8 5.64 17.14 1.86 0.35 0.64 0.16

exhibited moderate concentrations of N ­ H3-N, TP, F­ −, and best. This indicates that locations with satisfactory surface
LAS, high concentrations of DO, and the lowest concentra- water quality were characterized by enriched DO concentra-
tions of COD; the surface water quality of this group was the tions and organic pollutant and nutrient contents.

13
Environmental Earth Sciences (2021) 80:368 Page 11 of 16  368

Evaluation results of surface water quality (IWQI = 4.540). In addition, the IWQI method can be used
identification index to distinguish water pollution levels within groups classi-
fied as inferior to Class V.
Based on the water quality, the eight groups of samples
were classified according to the classes specified in the Spatial distribution of water quality
Environmental Quality Standards for Surface Water
(China 2002). One group (G6) belonged to Class III, two For a drainage basin located in an urban built-up area, espe-
groups (G7 and G8) belonged to Class IV, one group (G5) cially one that has a large number of factories and residential
belonged to Class V, and four groups (G1, G2, G3, and areas along both banks of the river, it is particularly impor-
G4) were classified as inferior to Class V (Table 8). The tant to gain an understanding of the temporal and spatial
surface water quality in 82.17% of the monitoring sec- variations in surface water quality to visualize the effects of
tions met the water quality target of the functional zones. human activities and tides on the water quality in the channel
The surface water quality was divided into unsuitable (estuaries). A geographic information system can be used to
(low quality) and suitable (acceptable quality) accord- obtain a comprehensive understanding of the spatial vari-
ing to the Environmental Quality Standards for Surface ability in surface water quality (Fig. 4).
Water (China 2002). The surface water quality in G5, G6, Figure 4a indicates that the surface water quality of the
G7, and G8 was classified as suitable, because it met the upper and middle reaches of the river is better than that of
threshold for the Class V standard (China 2002). In these the estuary, the surface water quality of the main stream is
groups, endogenous pollution and rainfall runoff were the better than that of the tributaries, and the water quality of the
dominant pollution sources, and the main background southwest tributaries is the worst in 2018. The IWQI results
pollutants for the water quality target of the functional of the all streams indicates that the surface water quality
zones were N ­ H3-N and TP. The surface water quality of falls short of the water quality target of the functional zone
G6 samples, which were collected from the upstream por- objectives, and the water quality is classified as “inferior
tion of the basin during the flood season in 2020, was the to Class V but not black and odorous”. The estuary water
best, and the water quality status was “good”. The surface pollution is more serious and the water quality is “inferior
water quality of four groups was classified as unsuitable, to Class V and black and odorous”. The main stream of the
because it did not meet the threshold for the Class V stand- river is fed by tributaries, and groundwater and rivers have
ard (China 2002). In these four groups, industrial sewage, a recharge–discharge relationship; therefore, the upper and
endogenous pollution, domestic sewage, and rainfall run- middle reaches of the main stream were less polluted than
off were the dominant pollution sources, and the primary the tributaries. During the nonflood season, the dilution and
background pollutants for the water quality target of the diffusion capacities of pollutants decrease as the flow of the
functional zones were COD, ­NH3-N, TP, and LAS. The river decreases, and this results in a decrease in the water
surface water quality of G1 samples, which were obtained quality. A comparison of the surface water quality during the
from the upper tributaries during the nonflood season in flood season, the nonflood season, and the annual average in
2018, was the worst and was classified as “inferior to Class 2018 indicated that the water quality during the flood season
V and black and odorous”. The IWQI method can be used was better than that during the annual average and the non-
to identify advantages and disadvantages among samples flood season. The key background pollutants for the water
of the same water quality grade. For example, the water quality target of the functional zones were COD, ­NH3-N, TP,
quality status of G7 and G8 is medium, but the water qual- DO, and LAS. External and internal pollution sources were
ity of G8 (IWQI = 4.060) is very good than that of G7 the major factors influencing the deterioration of the water

Table 8  The surface water No Individual water quality identification index Integrated water Rank
quality assessment results of 8 quality index
−
groups (unit: mg/L) DO COD NH3-N TP F LAS

G1 6.49 9.25 24.35 19.09 2.83 10.86 12.118 Inferior class V

G2 5.10 7.50 18.12 12.99 2.80 9.69 9.414 Inferior class V
G3 5.42 6.62 13.04 9.99 2.97 7.80 7.612 Inferior class V
G4 5.74 6.10 11.61 8.97 2.79 6.97 7.012 Inferior class V
G5 4.00 5.65 9.12 8.30 2.84 6.21 5.921 V
G6 3.08 4.70 5.17 3.62 2.58 2.63 3.650 III
G7 4.76 4.41 6.44 6.22 2.63 2.83 4.540 IV
G8 3.64 3.43 5.72 5.49 2.64 2.82 4.060 IV

13
 368   Page 12 of 16 Environmental Earth Sciences (2021) 80:368

Fig. 4  Surface water quality distribution map of Maozhou River Basin from 2018 to 2020

quality, and the main pollution sources were industrial sew- quality has clearly improved and the black-odorous water has
age, domestic garbage, rainfall runoff, and internal sources been eliminated. However, the water quality during nonflood
from a large number of factories (e.g., electroplating, printed season is still poor. This indicates that the water quality in
circuit board manufacturing, and optoelectronic device man- this basin has improved significantly over the past few years
ufacturing) and residential areas along both banks of the due to pollution treatment and management by the Chinese
river. In addition, during the early stages of rainfall, large government (e.g., thorough reforms, silt removal, and water
amounts of nonpoint source pollutants are transported into supply). The municipal sewage pipes along both banks of
the river, which increases the pollutant concentration; this the river have been decommissioned, and the industrial and
indicates that nonpoint source pollution has become the domestic sewage are collected and transported to a sewage
main factor affecting river water quality (Fragaa et al. 2016). treatment plant through the newly constructed municipal
Figure 4b indicates that spatial distribution of water qual- pipeline network. During the nonflood season, the accumu-
ity in 2019 was similar to that in the previous year. The water lation of surface pollutants causes severe nonpoint source

13
Environmental Earth Sciences (2021) 80:368 Page 13 of 16  368

pollution during rainfall (Ding et al. 2020). In addition, the reflected in the release of nutrients from sediments to the
lack of ecological base flow in the river leads to the alterna- overlying water, which was attributed to the disturbance of
tion of dry and wet sediment in the channel, which increases the sediment due to changes in water depth, and the release
the risk of pollutant release from the channel sediment of nutrients from the sediment to the overlying water under
(Yuan et al. 2020). After treatment at the sewage treatment anoxic conditions.
station, the tail water complies with discharge standards and The results demonstrated that the water quality status in
is then released into the river. The key background pollut- the Maozhou River basin has steadily improved during the
ants for the water quality target of the functional zones are monitoring period. The surface water quality at 82.17% of
­NH3-N, TP, and LAS. The main factors influencing river the monitoring sites met the water quality target of the func-
water quality are the unsuccessful decommissioning of sew- tional zones (surface water quality of Class V), with IWQI
age pipes, nonpoint source pollution induced by rainfall and values ranging from 12.157 to 3.650. The key sources of
endogenous pollution. During the nonflood season, the small pollutants in the surface water ecosystem in the study area
amount of water cannot effectively dilute the concentrations were organic and nutrient pollution from industrial sewage,
of nutrients and organic pollutants in the river, which leads domestic garbage, rainfall runoff, and internal sources. The
to the increase in the concentrations of the various pollutants main background pollutants for the water quality target of
and causes the deterioration of the water quality. the functional zones were ­NH3-N, TP, and LAS. Among
As the completion of various aquatic environment pollu- them, ­NH3-N and LAS were recognized as responsive and
tion control projects approaches, the improvement of surface sensitive to the surface water quality and spatio-temporal
water quality is becoming more significant in 2020 (Fig. 4c). variability.
The water quality in the main stream and the upstream tribu- From the perspective of time, the water quality of the
taries was “good”, while the water quality in the estuary river improved year by year. By 2019, the black-odorous
and the remaining tributaries remained stable at "medium to water had been eliminated. By 2020, the water quality of
good". The results showed that pollution occurs in the main the river had essentially met the water quality target of the
stream and that the pollution originates in the tributaries. functional zone objectives and the water quality status had
Furthermore, the results showed that the surface water qual- been steadily maintained at the “medium and good” level.
ity of the southwest tributaries and estuaries requires further From the perspective of space, the water quality of the
improvement. The surface water quality was significantly tributaries was worse than that of the main streams, while
influenced by ­NH3-N concentrations (p < 0.05), because the the surface water quality in the southeast tributaries was the
concentration of ­NH3-N during the nonflood season dem- worst. The surface water quality in the upstream portion of
onstrates a tendency to approach the water quality target of the main stream was better than that in the estuary. There-
the functional zones. fore, it is necessary to strengthen the water quality control of
the tributaries, especially the southeast tributaries.
A GIS suite that incorporated the results of multivariate
Conclusions statistical analysis and the IWQI has been proved to be suit-
able for the assessment of hydrochemical correlations, iden-
The spatio-temporal variation in the surface water quality in tification of pollution sources, and evaluation surface water
the Maozhou River basin from 2018 to 2020 was analyzed quality. This combined approach for water quality assess-
in this study. The surface water quality was evaluated using ment can be used to evaluate the surface water quality status
multivariate statistical techniques and the IWQI based on as well as the pollution levels of water samples that belong
the monthly average surface water quality data. The main to the same water quality rank. In addition, it can achieve the
conclusions are as follows: objective of quantitative evaluation of water quality that is
PCA was successfully applied to identify pollution inferior to Class V. This method is suitable for assessing sur-
sources. Three principal components defined during PCA face water quality at large scales, and it provides a scientific
were responsible for 83.1% of the total variance in the hydro- foundation for aquatic ecosystem management and planning,
chemical dataset, and three pollution sources were found to and helps to efficiently evaluate and manage surface water
be dominant, namely, domestic sewage and rainfall runoff, quality at river or basin scales. The methodology used in this
industrial sewage, and endogenous pollution. Domestic sew- study will serve as a reference for ecosystem projects in the
age and rainfall runoff mainly exhibited high concentrations Guangdong–Hong Kong–Macao Greater Bay Area that aim
of COD, TP, ­NH3-N, and LAS and low concentrations of to prevent surface water pollution.
DO, and were related to domestic sewage discharge and non-
point source pollution. Industrial sewage was characterized Acknowledgements  The authors extend particular thanks to the
Shenzhen municipal Ecology and Environment Bureau for provid-
by high concentrations of F– and was related to industrial ing water quality data. This study was supported by the project of
wastewater discharge. Endogenous pollution was mainly

13
 368   Page 14 of 16 Environmental Earth Sciences (2021) 80:368

Scientific Research Project of Science and Technology Innovation Fund Fang XQ, Fan Q, Li HW, Liao ZH, Xie JL, Fan SJ (2018) Multi-scale
from CCCC First Highway Consultants Co., Ltd (KCJJ2018-10, Key correlations between air quality and meteorology in the Guang-
Technologies of Comprehensive Regulation of Water Environment in dong–Hong Kong–Macau Greater Bay Area of China during
Highly Built Urban Areas). 2015–2017. Atmos Environ 191:463–477. https://​doi.​org/​10.​
1016/j.​atmos​env.​2018.​08.​018
Fragaa I, Charters FJ, Oullivan AD, Cochrane TA (2016) A novel
modelling framework to prioritize estimation of non-point
References source pollution parameters for quantifying pollutant origin
and discharge in urban catchments. J Environ Manag 167:75–84
Abbasi T, Abbasi SA (2012) Water quality indices. Elsevier, Gao YY, Qian H, Ren WH, Wang HK, Liu FX, Yang FX (2020)
Amsterdam Hydrogeochemical characterization and quality assessment of
Adimalla N, Qian H (2019) Hydrogeochemistry and fluoride contami- groundwater based on integrated-weight water quality index in a
nation in the hard rock terrain of central Telangana, India: analy- concentrated urban area. J Clean Prod. https://​doi.​org/​10.​1016/j.​
ses of its spatial distribution and health risk. SN Appl Sci 1:202. jclep​ro.​2020.​121006
https://​doi.​org/​10.​1007/​s42452-​019-​0219-​08 Grunsky EC, Drew LJ, Sutphin DM (2009) Process recognition in
Ana EMC, José ARL, Diego AMC, José TV, Josue DLB, Janete MR multi-element soil and stream-sediment geochemical data. Appl
(2019) Identification of the hydrogeochemical processes and Geochem 24:1602–1616. https://​doi.​org/​10.​1016/j.​apgeo​chem.​
assessment of groundwater quality, using multivariate statisti- 2009.​04.​024
cal approaches and water quality index in a wastewater irrigated Hamlat A, Guidoum A (2018) Assessment of groundwater quality in
region. Water 11:1702. https://​doi.​org/​10.​3390/​w1108​1702 the Seraidi region of Northwestern Algeria using water quality
Babiker IS, Mohamed MA, Hiyama T (2007) Assessing groundwater index (WQI). Appl Water Sci 8:220. https://​doi.​org/​10.​1007/​
quality using GIS. Water Resour Manage 21:699–715. https://d​ oi.​ s13201-​018-​0863-y
org/​10.​1007/​s1169-​006-​9059-6 Harvey ET, Kratzer S, Philipson P (2015) Satellite-based water qual-
Cai WB, Han BL, Lu F, Xian CF, Yang ZY (2020) Comprehensive ity monitoring for improved spatial and temporal retrieval of
evaluation of the eco-environment in the four global bay areas. chlorophyll-a in coastal waters. Remote Sens Environ 158:417–
Acta Ecol Sin 40:8392–8402. https://​doi.o​ rg/​10.​5846/​stxb2​02004​ 430. https://​doi.​org/​10.​1016/j.​rse.​2014.​11.​017
301062 Howladar MF, Al Numanbakth MA, Faruque MO (2017) An applica-
Chen L, He Q, Liu K, Li J, Jing C (2019) Downscaling of GRACE- tion of water quality index (WQI) and multivariate statistics to
derived groundwater storage based on the random forest model. evaluate the water quality around Maddhapara Granite Mining
Remote Sens. https://​doi.​org/​10.​3390/​rs112​42979 Industrial Area, Dinajpur. Bangladesh Environ Syst Res 6:1–18.
Chen YP, Yan H, Yao YJ, Zeng CL, Gao P, Zhuang LY, Fan LY, Ye DQ https://​doi.​org/​10.​1186/​s40068-​017-​0090-9
(2020) Relationships of ozone formation sensitivity with precur- Islam AT, Shen S, Haque MA, Bodrud-Doza M, Maw K, Habib MA
sors emissions, meteorology and land use types, in Guangdong- (2017) Assessing groundwater quality and its sustainability in
Hong Kong-Macao Greater Bay Area. China Environ Sci 94:1–13 Joypurhat district of Bangladesh using GIS and multivariate sta-
Chin YL, Mohd HA, Sarva MP, Aminatul HBY, Baba M (2012) Delin- tistical approaches. Environ Dev Sustain 20:1935–1959. https://​
eation of temporal variability and governing factors influencing doi.​org/​10.​1007/​s10668-​017-​9971-3
the spatial variability of shallow groundwater chemistry in a tropi- Jiang ZS, Li B, Wu JC, Tang T, Li H. Xu H (2019) Ecological effects
cal sedimentary island. J Hydrol. https://​doi.​org/​10.​1016/j.​jhydr​ of water environment management project on the Maozhou
ol.​2012.​02.​015 River Basin. J Hydroecol. https://​kns.​cnki.​net/​kcms/​detail/​42.​
China GB3838-2002 (2002) Environmental quality standards for sur- 1785.X.​20200​924.​1209.​002.​html
face water [S] Lambraski N, Antonakos A, Panagopoulos G (2004) The use of mul-
China HJ494-2009 (2009) Water quality guidance on sampling tech- ticomponent statistical analysis in hydrogeological environmen-
niques [S] tal research. Water Res 38:1862–1872. https://​doi.​org/​10.​1016/j.​
Dassi L (2011) Investigation by multivariate analysis of groundwater watres.​2004.​01.​009
composition in a multilayer aquifer system from North Africa: a Li BD, Liu C, Liu XB, Wang SY (2021) Water quality response char-
multi-tracer approach. Appl Geochem 26:1386–1398. https://​doi.​ acteristics and cause analysis of thermal stratification of large
org/​10.​1016/j.​apgeo​chem.​2011.​05.​012 reservoirs. J China Inst Water Resour Hydropow Res 19:1–9.
Ding FF, Liu YY, Wang L, Liu HF, Ji CY, Zhang L, Wu D (2020) https://​doi.​org/​10.​13244/j.​cnki.​jiwhr.​20200​113
Analysis of the palladium response relationship of a receiving Liang ZH, Chen XH, Luo H, Wu Q, Li MG (2020) Spatio-temporal
water body under multiple scenario changes in rainfall-runoff distribution characteristics of water quality in Shenzhen Bay
pollution. Environ Sci Pollut Res. https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​ and pollution source analysis. Water Resour Protect. https://​doi.​
s11356-​021-​12597-3 org/​10.​3880/j.​issn.​1004-​6933.​2020.​04.​015
Edegbene AO, Arimoro FO, Odume ON (2019) Developing and apply- Lin T, Xu PP, Qian H (2017) Assessment of water quality and analy-
ing a macroinvertebrate-based multimetric index for urban rivers sis of pollution source in Ningxia section of the Yellow River.
in the Niger Delta, Nigeria. Ecol Evol 9:12869–12885. https://d​ oi.​ Environ Chem 36:1388–1396
org/​10.​1002/​ece3.​5769 Liu YS, Tang YY, Zhong GS, Zeng H (2019) A comparison study on
Eilander D, Annor FO, Lannini L, Giesen NVD (2014) Remotely heavy metal/metalloid stabilization in Maozhou River sediment
sensed monitoring of small reservoir dynamics: a bayesian by five types of amendments. J Soils Sediments 19:3922–3933.
approach. Remote Sens 6:1191–1210. https://​doi.​org/​10.​3390/​ https://​doi.​org/​10.​1007/​s11368-​019-​02310-w
rs602​1191 Lou SH, Tang YD, Tao M, Ren KJ, Shao YH, Zhen WS (2020) Meth-
Emenike CP, Tenebe IT, Jarvis P (2018) Fluoride contamination in ods and practice of comprehensive improvement of Maozhou
groundwater sources in southwestern Nigeria: assessment using river water environment in Shenzhen, China. Water Wastew.
multivariate statistical approach and human health risk. Ecotox https://​doi.​org/​10.​19853/j.​zgjsps.​1000-​4602.​2020.​10.​001
Environ Safe 156:391–402. https://d​ oi.o​ rg/1​ 0.1​ 016/j.e​ coenv.2​ 018.​ Lv JJ, Yang JY, Liao WF, Jiang JH, Chen CH, Zhao J, Yang S (2014)
03.​022 Study on the water quality and environmental conditions of the
formation of black-odorous water. J Cent China Norm Univ

13
Environmental Earth Sciences (2021) 80:368 Page 15 of 16  368

(natural Sciences). https://​doi.​org/​10.​19603/j.​cnki.​1000-​1190.​ PM (2010) Global threats to human water security and river biodi-
2014.​05.​017 versity. Nature 467:555–561. https://d​ oi.o​ rg/1​ 0.1​ 038/n​ ature​ 09440
Madan KJ, Ankit SMAJ (2020) Assessing groundwater quality for Weng H, Kou J, Shao Q (2020) Evaluation of urban comprehen-
drinking water supply using hybrid fuzzy-GIS-based water quality sive carrying capacity in the Guangdong–Hong Kong–Macao
index. Water Res. https://​doi.​org/​10.​1016/j.​watres.​2020.​115867 Greater Bay Area based on regional collaboration. Environ
Mao QD, Xia D, Wen QP, Shi JZ (2019) A Revised Method of sur- Sci Pollut Res 27:20025–20036. https:// ​ d oi. ​ o rg/ ​ 1 0. ​ 1 007/​
face water quality evaluation based on background values and its s11356-​020-​08517-6
application to samples collected in Heilongjiang Province, China. Wu JG, Huang JH, Han XG, Gao XM, He FL, Jiang MX, Jiang ZG,
Water 11:1057. https://​doi.​org/​10.​3390/​w1105​1057 Primack RB, Shen ZH (2004) The three Gorges dam: an ecologi-
Mcgregor GB, Fabbro LD (2000) Dominance of cylindrospermopsis cal perspective. Front Ecol Environ 2:241–248. https://d​ oi.o​ rg/1​ 0.​
raciborskii (Nostocales, Cyanobacteria) in Queensland tropical 1890/​1540-​9295(2004)​002[0241:​TTGDAE]​2.0.​CO;2
and subtropical reservoirs: implications for monitoring and man- Wu J, Zhou H, He S, Zhang Y (2019) Comprehensive understanding
agement. Lakes Reserv Res Manag 5:195–205. https://d​ oi.o​ rg/1​ 0.​ of groundwater quality for domestic and agricultural purposes in
1046/j.​1440-​1770.​2000.​00115x terms of health risks in a coal mine area of the Ordos basin, north
Naveedullah N, Hashmi MZ, Yu CN, Shen CF, Muhammad N, Shen of the Chinese Loess Plateau. Environ Earth Sci 78:1–17. https://​
H, Chen YX (2016) Water quality characterization of the Siling doi.​org/​10.​1007/​s12665-​019-​8471-1
reservoir (Zhejiang, China) using water quality index. Clean: Soil, Xiao J, Wang LQ, Deng L, Jin ZD (2019) Characteristics, sources,
Air, Water 44:553–562. https://​doi.​org/​10.​1002/​clen.​20140​0126 water quality and health risk assessment of trace elements in river
Nong XZ, Shao DG, Zhong H, Liang JK (2020) Evaluation of water water and well water in the Chinese Loess Plateau. Sci Total Envi-
quality in the South-to-North Water Diversion Project of China ron. https://​doi.​org/​10.​1016/j.​scito​tenv.​2018.​09.​322
using the water quality index (WQI) method. Water Res. https://​ Xu ZX (2005) Comprehensive water quality identification index for
doi.​org/​10.​1016/j.​watres.​2020.​115781 environmental quality assessment of surface water. J Tongji Univ
Peyman N, Sany SBT, Tajfard M, Hashim R, Rezayi M, Karlen DJ (natural Science) 33:483–488
(2017) The status and characteristics of eutrophication in tropical Xu PP, Zhang QY, Qian H, Li MN, Hou K (2019a) Characterization
coastal water. Environ Sci Process Impacts 19:1086–1103. https://​ of geothermal water in the piedmont region of Qinling Mountains
doi.​org/​10.​1039/​c7em0​0200a and Lantian-Bahe Group in Guanzhong Basin, China. Environ
Ramos MAG, de Oliveira ESB, Piao ACS, Leite DAND, de Angelis Earth Sci 78:1–7. https://​doi.​org/​10.​1007/​s12665-​019-​8418-6
DD (2016) Water quality index (WQI) of Jaguari and Atibaia Xu PP, Feng WW, Qian H, Zhang QY (2019b) Hydrogeochemical
rivers in the region of Paulinia, Sao paulo. Brazil Environ Monit characterization and irrigation quality assessment of shallow
Assess. https://​doi.​org/​10.​1007/​S10661-​016-​5261-z groundwater in the Central-Western Guanzhong Basin, China.
Sany SBT, Monazami G, Rezayi M, Tajfard M, Borgheipour H (2019) Int J Environ Res Public Health 16:18. https://​doi.​org/​10.​3390/​
Application of water quality indices for evaluating water quality ijerp​h1609​1492
and anthropogenic impact assessment. Int J Environ Sci Technol Xu PP, Li MN, Qian H, Zhang QY, Liu FX, Hou K (2019c) Hydro-
16:3001–3012. https://​doi.​org/​10.​1007/​s13762-​018-​1894-5 chemistry and geothermometry of geothermal water in the central
Shi W, Xia J, Zhang X (2016) Influences of anthropogenic activities Guanzhong Basin, China: a case study in Xi’an. Environ Earth
and topography on water quality in the highly regulated Huai Sci. https://​doi.​org/​10.​1007/​s12665-​019-​8099-1
River basin, China. Environ Sci Pollut Res 23:21460–21474. Xu PP, Zhang QY, Qian H, Li MN, Yang FX (2021) An investigation
https://​doi.​org/​10.​1007/​s11356-​016-​7368-8 into the relationship between saturated permeability and micro-
Shi XY, Wang Y, Jiao JJ, Zhong JL, Wen HG, Dong R (2018) Assess- structure of remolded loess: a case study from Chinese Loess
ing major factors affecting shallow groundwater geochemical Plateau. Geoderma. https://d​ oi.o​ rg/1​ 0.1​ 016/J.G
​ EODER ​ MA.2​ 020.​
evolution in a highly urbanized coastal area of Shenzhen City, 114774
China. J Geochem Explor 184:17–27. https://​doi.​org/​10.​1016/j.​ Yuan HZ, Yin HB, Yang Z et al (2020) Diffusion kinetic process of
gexplo.​2017.​10.​003 heavy metals in lacustrine sediment assessed under different redox
Shi GJ, Guo LY, Lin T, Zhang YY (2020) Variation characteristics conditions by DGT and DIFS model. Sci Total Environ. https://​
and future changes of precipitation in Shenzhen. J Yangtze River doi.​org/​10.​1016/j.​scito​tenv.​2020.​140418
Sci Res Inst 37:28–33. https://​doi.​org/​10.​11988/​ckyyb.​20190​088 Zhang XD, Qian H, Chen J, Qiao L (2014) Assessment of groundwa-
Shyu GS, Cheng BY, Chiang CT, Yao PH, Chang TK (2011) Applying ter chemistry and status in a heavily used semi-arid region with
factor analysis combined with Kriging and information entropy multivariate statistical analysis. Water 6:2212–2232. https://​doi.​
theory for mapping and evaluating the stability of groundwater org/​10.​3390/​w6082​212
quality variation in Taiwan. Int J Environ Res Pub Health 8:1084– Zhang QY, Xu PP, Qian H (2019) Assessment of groundwater quality
1109. https://​doi.​org/​10.​3390/​ijerp​h8041​084 and human health risk (HHR) evaluation of nitrate in the Cen-
Sun R, Luo X, Tang B, Li Z, Wang T, Tao L, Mai B (2016) Persis- tral-Western Guanzhong Basin, China. Int J Environ Res Public
tent halogenated compounds in fish from rivers in the Pearl River Health. https://​doi.​org/​10.​3390/​ijerp​h1621​4246
Delta, South China: geographical pattern and implications for Zhang YS, Wu L, Ren HZ, Deng LC, Zhang PC (2020c) Retrieval of
anthropogenic effects on the environment. Environ Res 146:371– water quality parameters from hyperspectral images using hybrid
378. https://​doi.​org/​10.​1016/j.​envres.​2016.​01.​021 bayesian probabilistic neural network. Remote Sens 12:1567.
Varol M (2019) Arsenic and trace metals in a large reservoir: seasonal https://​doi.​org/​10.​3390/​rs121​01567
and spatial variations, source identification and risk assessment Zhang QY, Xu PP, Qian H (2020b) Groundwater Quality Assessment
for both residential and recreational users. Chemosphere 228:1–8. Using Improved Water Quality Index (WQI) and Human Health
https://​doi.​org/​10.​1016/j.​chemo​sphere.​2019.​04.​126 Risk (HHR) evaluation in a semi-arid region of Northwest China.
Varol M, Gokot B, Bekleyen A, Sen B (2012) Spatial and temporal Expo Health. https://​doi.​org/​10.​1007/​s12403-​020-​00345-w
variations in surface water quality of the dam reservoirs in the Zhang QY, Xu PP, Qian H, Yang FX (2020a) Hydrogeochemistry and
Tigris River basin, Turkey. CATENA 92:11–21. https://​doi.​org/​ fluoride contamination in Jiaokou Irrigation District, Central
10.​1016/j.​catena.​2011.​11.​013 China: assessment based on multivariate statistical approach and
Vorosmarty CJ, McIntyre PB, Gessner MO, Dudgeon D, Prusevich A, human health risk. Sci Total Environ. https://​doi.​org/​10.​1016/j.​
Green P, Glidden S, Bunn SE, Sullivan CA, Liermann CR, Davies scito​tenv.​2020.​140460

13
 368   Page 16 of 16 Environmental Earth Sciences (2021) 80:368

Zheng G (2010) Investigation and assessment on heavy metals pollu- Publisher’s Note Springer Nature remains neutral with regard to
tion of agricultural soil in Guanzhong Irrigation District. Chin jurisdictional claims in published maps and institutional affiliations.
J Soil Sci 41:473–478. https://​doi.​org/​10.​19336/j.​cnki.​trtb.​2010.​
02.​044
Zhou H, Shi P, Wang JA, Yu D, Gao L (2010) Rapid urbanization and
implications for river ecological services restoration: case study
in Shenzhen, China. J Urban Plan Dev 137:121–132. https://​doi.​
org/​10.​1061/​(ASCE)​UP.​1943-​5444.​00000​51

13