Desalination and Water Treatment 239 (2021) 137–149

November
www.deswater.com
doi: 10.5004/dwt.2021.27715

Evaluation of spatial and seasonal water quality variation of urban lake by

multivariate statistical approaches for water quality improvement

Siping Niua,*, Xiaolong Songa, Jianghua Yub, Jing Wua

a
Department of Environmental Science and Engineering, School of Energy and Environment, Anhui University of Technology,
Ma’anshan 243002, China, emails: [email protected] (S. Niu), [email protected] (X. Song), [email protected] (J. Wu)
b
Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, Jiangsu Key Laboratory of Atmospheric
Environment Monitoring and Pollution Control (AEMPC), Nanjing University of Information Science & Technology, Nanjing 210044,
China, email: [email protected]

Received 5 March 2021; Accepted 9 August 2021

abstract
In addition to water volume, quality is also an important factor affecting the use of water resources,
which is greatly related to the development of the modern city. This study was conducted to obtain
information on spatial-temporal distribution and major sources of pollutants in an urban lake
(Yushan Lake, Ma’anshan City, China). The integrated multivariate statistical approaches, includ-
ing Pearson correlation analysis, cluster analysis (CA) and principal component analysis (PCA)
were utilized to analyze the lake water quality in spring, summer, autumn and winter during 2019.
CA revealed the spatial and seasonal variation of lake water quality as well as the correlated param-
eters. PCA provided the pollution sources via figuring out the predominant factors explaining over
70% of the total variation of the variables. Based on the result of the present study, those sites with
significant water pollution as well as the corresponding parameters were identified. Subsequently,
the strategy, against specific pollution source was offered to improve the lake water quality.
Keywords: Urban lake; Water quality; Multivariate statistical analysis; Comprehensive evaluation;
Spatial-temporal variation

1. Introduction been degraded in many places over the world [14,15]. A

large number of studies have been carried out to focus on
Surface water, as the major water body, plays a vital role
lake water quality. Wu et al. [16] assessed the water qual-
in maintaining ecological stability and supporting socio-
ity of Shahu Lake from China and suggested ways, includ-
economic development [1–4]. However, due to various
ing circulation accelerating and lake water replenishment,
human activities throughout urbanization and moderniza-
which could be taken to improve the water quality. Wang
tion, it has been vulnerable to pollution [5–7]. In general,
et al. [17] evaluated the eutrophication and water qual-
both natural processes and anthropogenic activities can
ity in the estuarine area of Lake Wuli (China) and found
impair the quality of surface water [8–10]. The discharge
the east part of Wuli Lake had worse eutrophication lev-
of wastewater/runoff, which is accompanied by industrial
els and water quality status than other parts. Davraz et al.
and agricultural activities, is confirmed as a significant
[18] analyzed the relationship between water quality and
contributor to surface water quality degradation [11–13].
hydrogeochemical processes of Salda alkaline lake (Burdur,
Lake is usually in the service of irrigation, recreation,
Turkey), observing the water chemical structure was mainly
fishing and even source water. However, its quality has
controlled by the weathering processes and anthropogenic

* Corresponding author.

1944-3994/1944-3986 © 2021 Desalination Publications. All rights reserved.

138 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

inputs. Quan et al. [15] studied the mechanism of anthro- annual average temperature and precipitation are 16°C and
pogenic actives on water quality change of Baiyangdian 1,100 mm, respectively. Yushan Lake, the largest urban lake
Lake in North China between 2006 and 2016 and provided of Ma’anshan City, from the downtown area is an import-
a more thorough understanding of the driving mechanism ant landscape. Specifically, the lake has an area of 132.47 ha
of water quality change. Those studies on lake water quality with a perimeter of 9.87 km and an average water depth
are the potential to benefit lake water quality improvement of 1.2 m, resulting in a water capacity of 829,800 m3. Water
and lake ecological function restoration of a specific area. volume is controlled pronouncedly by the land-based
To evaluate the water quality, regular monitoring is runoff. Yushan Lake connects with the Yangtze River
carried out frequently for water environment management via the Yushan River and the Yongfeng River.
[19,20]. In general, a complex parameter matrix, including Surface water from a depth of ~20 cm of Yushan Lake
designative variables, is available. However, it is uneasy was sampled in acid-cleaned, 2 L plastic bottles that were
to distinguish the most meaningful information compre- rinsed with surface water beforehand. Data sets consisting
hensively attributed to the Spatio-temporal variation of of 13 variables from 16 monitoring sites in spring, summer,
monitoring data among sampling sites [21]. As the credible autumn and winter of 2019 were available for the present
alternative, the statistical analysis, represented by multivar- study (Fig. 1).
iate statistical analysis (MSA), has been used for water qual-
ity assessment of river water, groundwater, lake water and 2.2. Water quality analysis
mine water [22–27].
MSA can process a large volume of monitoring data Water temperature (WT), pH, electroconductibility
from numerous sites. We can group the monitoring stations (EC) and dissolved oxygen (DO) were measured in situ
based on their similarity of water quality and apportion the by using Multiparameter Water Quality Sonde Method.
pollution sources of regional water bodies [28–30]. MSA To ensure the accuracy of monitoring results the sensors
technique mainly consists of cluster analysis (CA), prin- were calibrated carefully prior to the test. Other indicators
cipal component analysis (PCA), factor analysis (FA) and comprising chemical oxygen demand (COD), total sus-
discriminant analysis (DA) [31,32], among which CA and pended solids (TSS), total nitrogen (TN), total dissolved
PCA are the most popular [33,34]. In order to obtain accu- nitrogen (DTN), ammonium (NH4–N), nitrate (NO3–N),
rate information, several methods are selected simultane- total phosphorus (TP), total dissolved phosphorus (DTP)
ously as the MSA technique is used in practice. Sebastián and orthophosphate (PO4–P) were analyzed in the labora-
Gradilla-Hernández et al. [34] using PCA, CA and DA ana- tory within 48 h after sampling. Filtered samples by GF/F
lyzed the water quality of Lake Cajititlán, Mexico, giving the filters (Whatman, Kent Great Britain) were used to deter-
temporal behavior of water quality associated with seasonal mine DTN, NH4–N, NO3–N, DTP and PO4–P. The determi-
transport pathways of pollutants. Via and CA, DA, PCA and nation of water quality parameters was carried out based
FA, Quan et al. [15] assessed the temporal and spatial varia- on those methods summarized in Table 1.
tions of water quality and figured out the dominant factors
of the variations selected in Baiyangdian Lake. Hatvani et 2.3. Data analysis
al. [35] investigated the Spatio-temporal changes and driv-
We extracted lake water quality information by integrat-
ers of trophic status over three decades in the largest shal-
ing multivariable analysis consisting of Pearson correlation
low lake in Central Europe by using combined cluster and
analysis, PCA and CA. PCA associated Pearson correlation
DA and exploratory PCA. MSA has been proved as a vital
was used to interpret the relationship among variables as
technology in regional water environment management.
well as the pollution sources, while CA for spatiotemporal
Recently, the Chinese government is sparing no effort
variation of lake water quality. PCA and CA were oper-
to improve the quality of freshwater with tremendous res-
ated by using the data through Z-scale transformation to
toration on lakes and rivers [36]. Under this condition, the
minimize the influence on the classification due to variable
water quality evaluation on the surface water environment
size and measurement unit. All mathematical and statisti-
is of critical importance. This research towards an urban
cal analyses were done by using Microsoft Excel (Version
lake from the Yangtze River Basin was aimed to (1) reveal
2016) and SPSS (Version 20.0).
the water quality indicator responsible for spatial and tem-
poral variation, and (2) identify both the main influencing
factors of water quality and the pollution sources and (3) 3. Results and discussion
provide suggestion on water quality management and
3.1. Water quality evaluation
improvement.
Temperature, DO, pH and EC are commonly basic
parameters significantly related to the quality of surface
2. Materials and methods water. SS, COD, nitrogen and phosphorus are typical indi-
cators of water pollution by specific substances. 13 Variables
2.1. Study area and sampling
involved in the present study are summarized in Figs. 2–4.
Ma’anshan City, covering 4,049 km2 with a popula- The result of the Pearson correlation coefficient among 13
tion of ~2.30 million, is located in the lower reaches of parameters is given in Table 2.
the Yangtze River, China. Within the subtropical mon- WT can influence various physical, chemical and bio-
soon zone, the city has hot and rainy weather in sum- chemical processes, causing significant environmental
mer while mild and dry in winter. It is recorded that the impacts [32]. In this study, WT changed from 8.8°C to 33.7°C.
 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149 139

Fig. 1. Sampling sites.

Table 1
Water quality parameters measured and relevant analytical methods

Variables Analytical methods Instruments

Water temperature (WT), °C Multiparameter Water Quality SX725 DO/pH Meter (Shanghai San-Xin
Sonde Instrumentation, Inc., Shanghai, China)
pH Multiparameter Water Quality SX725 DO/pH Meter (Shanghai San-Xin
Sonde Instrumentation, Inc., Shanghai, China)
Dissolved oxygen (DO), mg L–1 Multiparameter Water Quality SX725 DO/pH Meter (Shanghai San-Xin
Sonde Instrumentation, Inc., Shanghai, China)
Electroconductibility (EC), µS cm–1 Multiparameter Water Quality X751 Conductivity Meter (Shanghai San-Xin
Sonde Instrumentation, Inc., Shanghai, China)
Total suspended solids (TSS), mg L–1 Dryness at 103°C–105°C XGQ-2000 Electric Blast Drying Oven, (Yuyao Star
Instrument Factory, Yuyao, China)
Chemical oxygen demand (COD), mg L–1 Closed Reflux, Titrimetric HACH DRB200 COD Digestion Device
Method (Hach China, Shanghai, China)
Total nitrogen (TN), mg L–1 Potassium Persulfate oxidation TU-1901 UV-Vis Spectrophotometer (Beijing Purkinje
General Instrument Co., Ltd., Beijing, China)
Total dissolved nitrogen (DTN), mg L–1 Potassium Persulfate oxidation TU-1901 UV-Vis Spectrophotometer (Beijing Purkinje
General Instrument Co., Ltd., Beijing, China)
Ammonium (NH4–N), mg L–1 Nesslerization L6S UV-Visible Spectrophotometer (INESA
Analytical Instrument Co., Ltd., Shanghai, China)
Nitrate (NO3–N), mg L–1 Phenol Disulfonic Acid TU-1901 UV-Vis Spectrophotometer (Beijing Purkinje
Spectrophotometric General Instrument Co., Ltd., Beijing, China)
Total phosphorus (TP), mg L–1 Potassium Persulfate oxidation L6S UV-Visible Spectrophotometer (INESA
Analytical Instrument Co., Ltd., Shanghai, China)
Total dissolved phosphorus Potassium Persulfate oxidation L6S UV-Visible Spectrophotometer (INESA
(DTP), mg L–1 Analytical Instrument Co., Ltd., Shanghai, China)
Orthophosphate (PO4–P), mg L–1 Ascorbic Acid Method L6S UV-Visible Spectrophotometer (INESA
Analytical Instrument Co., Ltd., Shanghai, China)
140 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

35 10
(a) (b)

30
9

25
8

DO (mg L )
-1
WT (°C)

20

7
15

6
10

5 5
Spring Summer Autumn Winter Spring Summer Autumn Winter
9.0 360
(c) (d)
340
8.5

320
8.0
EC (s cm )
-1

300
7.5
pH

280

7.0
260

6.5
240

6.0 220
Spring Summer Autumn Winter Spring Summer Autumn Winter

Fig. 2. Water temperature, DO, pH and EC of lake water.

70 70
(a) (b)
60 60

50 50
COD (mg L )
-1
SS (mg L )

40 40
-1

30 30

20 20

10 10

0 0
Spring Summer Autumn Winter Spring Summer Autumn Winter

Fig. 3. SS and COD of lake water.

As expected, it had a remarkably seasonal variation with pH values were observed to vary between 7.10 and 8.52,
the highest in summer while the lowest in winter. WT had suggesting the lake water was alkaline, especially during
significant relationship with SS, DTN and TP (r = 0.723 spring. The high pH was detected at Sites 11–15 in spring
for SS, 0521 for DTN and 0.824 for TP, p < 0.05) in summer and at Sites 1–4 in summer. Meanwhile, pH had significantly
while with pH during autumn (r = 0.538, p = 0.032). positive relationship with both WT and SS (r = 0.270 for
 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149 141

3.5 3.5
(a) (b)
3.0 3.0

2.5 2.5

DTN (mg L )
TN (mg L )

-1
2.0 2.0
-1

1.5 1.5

1.0 1.0

0.5 0.5

0.0 0.0
Spring Summer Autumn Winter Spring Summer Autumn Winter
3.5 3.5
(c) (d)
3.0 3.0

2.5 2.5
NH4 -N (mg L )

NO3 -N (mg L )
-1

-1

2.0 2.0

1.5 1.5

1.0 1.0

0.5 0.5

0.0 0.0
Spring Summer Autumn Winter Spring Summer Autumn Winter
0.20 0.05
(e) (f)

0.04
0.15

0.03
DTP (mg L )
-1
TP (mg L )
-1

0.10

0.02

0.05
0.01

0.00 0.00
Spring Summer Autumn Winter Spring Summer Autumn Winter
0.05
(g)

0.04
PO4 -P (mg L )

0.03
-1

0.02

0.01

0.00
Spring Summer Autumn Winter

Fig. 4. Nutrients of lake water.

142 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

WT and r = 6.02 for SS, p < 0.05). This indicates that pH and Fig. 4e–g shows the phosphorus present in the lake water.
TSS might be affected by the algae growth, related to water Different from nitrogen, particle-combined phosphorus was
temperature. Otherwise, Barakat et al. [32] found pH was considerable, which accounted for 44.44%–91.67% with an
negatively correlated with SS in river water. average of 73.61% to the total phosphorus in mass. In sum-
DO concentrations of Yushan Lake appeared with mer, TP concentration was the highest due to the contributor
6.36–9.85 mg L–1 in spring, 5.14–6.97 mg L–1 in summer, of non-point source pollution.
5.70–8.70 mg L–1 in autumn and 6.46–7.54 mg L–1 in whiter, Statistical analysis on those parameters involved in the
also having an obvious seasonal variation. Specifically, the Chinese Surface Water Quality Guidelines (GB3838-2002) is
lowest DO appears at site 2 (5.14 mg L–1) in summer, while summarized in Table 3. Despite the strong seasonal varia-
the highest at Site 15 (9.85 mg L–1) in winter. The lowest tion, the values of pH still were in the range of 6–9, which
DO level of lake water in summer was attributed to either is required by the Chinese standard. DO has the mini-
increased temperature reducing the dissolution of oxygen mum level of 5.19 mg L–1 in summer while the maximum
in water or intensive microorganisms’ activity consuming of 9.85 mg L–1 in spring, coupled with water quality for the
large amounts of DO for organic matter degradation or worst-case as class III. Based on averaged COD concentra-
pollutant transformation [37,38]. tion, the lake water quality can be considered as class IV or
EC was observed to have the highest level in V. Both during spring and autumn, TN had a concentration
winter (338.56 ± 8.45 µS cm–1) followed by spring less than 1.0 mg L–1. Therefore, overall, the lake water can be
(326.50 ± 2.78 µS cm–1), autumn (288.81 ± 2.88 µS cm–1) and classified as grade III in spring and IV in autumn. While, in
summer (238.75 ± 1.64 µS cm–1). Meanwhile, it showed an winter the concentration of TN was over 2.0 mg L–1, going
obviously negative correlation with COD while positive beyond the limit of class V. During four seasons, NH4–N
with NH4–N and NO3–N. changed between 0.13 and 0.94 mg L–1, suggesting the water
SS concentration had a large range between 1 and quality class belonged to I to III. Based on TP, the water
50 mg L–1, with a mean of 31 ± 8 mg L–1 for spring, 10 ± 5 mg L–1 quality varied between IV and V depending on sites. On
for summer, 22 ± 6 mg L–1 for autumn and 6 ± 4 mg L–1 for the whole, the water quality of Yushan Lake was class IV
winter (Fig. 3a). Attributed to low temperatures, the growth in spring, summer and autumn, while poor V in winter.
of algae was inhibited in winter. Otherwise, in spring with
increasing temperature both the growth of algae, causing 3.2. Spatial similarity and site grouping
high SS level. Though algae biomass was significantly higher
in summer than in other seasons, daily salvaging for algae CA is very helpful in sorting out the spatial group based
reduced SS concentration to some degree. As to autumn, on the water quality similarity of sampling sites. The sites
plant litterfall entering the lake increased SS level again. from one group have similar water quality. The results of
As the representative of organic matter pollution, COD CA by Ward’s methods are available in Fig. 5. For Yushan
had values of 3.92 to 71.63 mg L–1, with evidently seasonal Lake, 16 sites were grouped into three clusters regardless
variation. Specifically, the highest occurred in summer of the season. Those pollutants present in lake water were
especially at Sites 5, 7, 12 and 15. Those sites are located derived from external sources (land-based) and internal
near the shore areas, where stormwater was discharged sources (sediments). The migration of pollutants from
directly. On the other hand, COD had a native relationship sediments was controlled by hydrodynamic and physico-
with SS, especially in spring (r = –0.403), which means that chemical conditions of the water environment. The land-
organic matter mainly existed in dissolved forms in the lake. based sources, like stormwater, also could not be persistent.
Nitrogen with various species is given in Fig. 4a–d. Therefore, the contribution of pollution sources varied
TN, DTN, NH4–N and NO3–N had the concentrations with season and some sites were grouped into different
of 0.77–1.28 mg L–1, 0.42–0.77 mg L–1, 0.18–0.60 mg L–1 groups in different seasons. The statistical analysis of water
and 0.06–0.11 mg L–1 in spring, 0.90–2.01 mg L–1, 0.50– quality data depending on the cluster is given in Table 4.
0.70 mg L–1, 0.17–0.39 mg L–1 and 0.22–0.28 mg L–1 in sum- In spring, with the increased water temperature, the
mer, 1.02–1.33 mg L–1, 0.54–0.81 mg L–1, 0.13–0.21 mg L–1 biological activity got intensive. Cluster 1 consisted of
and 0.18–0.50 mg L–1 in autumn, and 2.14–3.15 mg L–1, Sites 2, 3, 5 and 9, from the stagnant water regions close
1.84–2.90 mg L–1, 0.36–0.94 mg L–1 and 1.14–1.56 mg L–1 in to the shore with the highest concentrations of COD and
winter. Surprisingly, they all had the highest level in win- DTN, which was related to the leaching of materials from
ter. Furthermore, TN, DTN and NO3–N occurred with the the plant litter (leaves). Cluster 2 was grouped by Sites 1,
lowest level in spring. It was found that dissolved nitro- 4, 6–8, 10, 14 and 15, mainly coming from the places with
gen was the dominant species with the proportion to total frequent aquatic activity, coupled with obviously higher val-
nitrogen as 56.57% ± 7.97%, 55.44% ± 9.15%, 59.85% ± 8.21% ues of NH4–N than Cluster 1. This result is likely due to the
and 88.77% ± 4.23% in spring, summer, autumn and winter, washing of mop/clothes of local inhabitants as well as the
respectively. The level of nitrogen, especially for dissolved release from sediments caused by hydrodynamic factors.
forms, is usually regulated by biological processes. It was Cluster 3 included Sites 11–13, 16, having the highest pH,
found that the water temperature was significantly nega- SS and TP. Those sites were within the zone connecting with
tively correlated with TN, DTN, NH4–N and NO3–N in this several ponds, which had bad water quality, near Golden
study. So, the conversion/reduction of nitrogen in the lake Eagle Business Circle. It represents the effects of the human
during winter was depressed by the low temperature. And activities from Yushan Park and business areas on lake
this is one of the reasons for the lake with the highest nitro- water. And the release inside the lake is believed to be the
gen level in winter. major factor governing the water quality in this season.
Table 2
Pearson correlation coefficient among 13 parameters

WT pH DO EC SS COD TN DTN NH4–N NO3–N TP DTP PO4–P

WT 1.000
pH 0.270* 1.000
DO –0.096 0.228 1.000
EC –0.797** –0.074 0.524** 1.000
SS 0.228 0.602** 0.592** 0.199 1.000
COD 0.365** –0.046 –0.23 –0.333** –0.11 1.000
TN –0.787** –0.547** –0.155 0.516** –0.559** –0.183 1.000
DTN –0.829** –0.551** –0.139 0.578** –0.543** –0.189 0.972** 1.000
NH4–N –0.573** –0.383** 0.204 0.663** –0.215 –0.192 0.714** 0.759** 1.000
NO3–N –0.825** –0.565** –0.216 0.509** –0.612** –0.213 0.961** 0.982** 0.690** 1.000
TP 0.601** 0.016 0.179 –0.324** 0.273* 0.155 –0.250* –0.296* –0.028 –0.347** 1.000
DTP –0.274* –0.291* –0.019 0.227 –0.317* 0.021 0.498** 0.477** 0.478** 0.434** 0.025 1.000
PO4–P –0.043 –0.357** –0.289* –0.08 –0.504** 0.186 0.458** 0.421** 0.370** 0.413** 0.082 0.625** 1.000

*correlation is significant at the 0.05 level (2–tailed);

**correlation is significant at the 0.01 level (2–tailed).

Table 3
Statistical analysis on the parameters involved in Chinese Surface Water Quality Guidelines (GB3838-2002)

Class WT (°C) pH DO (mg L–1) COD (mg L–1) TN (mg L–1) NH4–N (mg L–1) TP (mg L–1)
Spring 23.5–25.6 (24.6) 7.18–8.52 (8.02) 6.36–9.85 (8.55) 7.28–71.63 (24.59) 0.77–1.28 (0.96) 0.18–0.60 (0.37) 0.079–0.148 (0.100)
Summer 32.3–33.7 (32.7) 7.23–7.91 (7.66) 5.19–6.97 (6.14) 22.53–39.44 (32.75) 0.90–2.01 (1.12) 0.17–0.39 (0.22) 0.088–0.140 (0.108)
Autumn 20.0–20.6 (20.3) 7.24–8.41 (8.09) 5.70–8.70 (6.83) 11.20–33.60 (25.34) 1.02–1.33 (1.12) 0.13–0.21 (0.17) 0.054–0.077 (0.066)
S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

Winter 8.8–9.4 (9.1) 7.10–7.66 (7.42) 6.46–7.54 (6.92) 3.92–48.72 (20.23) 2.14–3.15 (2.47) 0.36–0.94 (0.55) 0.054–0.095 (0.074)
Week increasing in ≤1 6–9 ≥7.5 ≤15 ≤0.2 ≤0.15 ≤0.01
I
Week decreasing in ≤2
II 6–9 6–7.5 ≤15 0.2–0.5 0.15–0.5 0.01–0.025
III 6–9 5–6 15–20 0.5–1.0 0.5–1.0 0.025–0.05
IV 6–9 3–5 20–30 1.0–1.5 1.0–1.5 0.05–0.1
V 6–9 2–3 30–40 1.5–2.0 1.5–2.0 0.1–0.2
143
 144 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

0.009 ± 0.005 In the downtown area, the control of stormwater has

0.007 ± 0.003
0.008 ± 0.002

0.018 ± 0.006
0.012 ± 0.005
0.007 ± 0.004

0.005 ± 0.001
0.010 ± 0.004
0.007 ± 0.002

0.012 ± 0.002
0.012 ± 0.003
0.022 ± 0.004
not been put into practice. During summer the influence of
stormwater significantly took place. Cluster 1 includes Sites
2–6, 8, 9 and 11, receiving the road runoff from Yushan Park.
PO4–P

Those sites had the highest DTP and PO4–P but the lowest
DO concentration. Cluster 2 with Sites 1, 7, 10, 12, 13, corre-
sponding to the lowest DTN level and medium pollution by
0.019 ± 0.004
0.022 ± 0.005
0.019 ± 0.002

0.024 ± 0.006
0.019 ± 0.004
0.018 ± 0.001

0.017 ± 0.004
0.018 ± 0.005
0.022 ± 0.003

0.019 ± 0.005

0.035 ± 0.007
0.026 ± .006
PO4–P. While Cluster 3 grouped by Sites 14–16 was related
to the discharge of runoff originated from traffic roads. As
DTP

a result, the highest levels of SS, DTN, NH4–N and TP were

observed.
For autumn, Cluster 1 consisted of Sites 1, 5, 6, 8, 9,
0.085 ± 0.006
0.093 ± 0.009
0.130 ± 0.014

0.100 ± 0.010
0.106 ± 0.006
0.136 ± 0.003

0.067 ± 0.007
0.058 ± 0.004
0.069 ± 0.007

0.073 ± 0.011
0.066 ± 0.012
0.089 ± 0.006
10–12, which are those places with high levels of TN and
TP. This was attributed to the soaking of plant litters falling
into the lake. Cluster 2 comprised Sites 2–4, arising with the
TP

highest pH, NO3–N and PO4–P while the lowest COD, TN

and TP. This cluster shows the water quality of the northern
0.10 ± 0.01
0.09 ± 0.01
0.07 ± 0.01

0.27 ± 0.01
0.22 ± 0.00
0.25 ± 0.01

0.27 ± 0.05
0.36 ± 0.10
0.20 ± 0.02

1.20 ± 0.03
1.24 ± 0.04
1.55 ± 0.01
part with slow water exchange. Cluster 3 containing Sites
NO3–N

7, 13–16 gave the water quality information of those places,

with medium contamination by COD, TN and PO4–P. It can
be found that the water quality of the southern part was
homogeneous during autumn.
0.23 ± .0.04

0.21 ± 0.03
0.21 ± 0.04
0.25 ± 0.10

0.18 ± 0.02

0.15 ± 0.01

0.89 ± 0.04
0.31 ± .07
0.47 ± .07

0.19 ± .01

0.43 ± .03
0.45 ± .06
Statistical analysis on water quality data depending on cluster (Mean ± STD; Unit: WT, °C; pH, --; others, mg L–1)

As to winter, the sites were also classified into 3 clus-

NH4–N

ters due to the water quality. Sites 7, 8, 11, 13 and 15 from

Cluster 1 represented the water quality from shallow zones
with less pollution by nitrogen. Otherwise, Cluster 2 is com-
0.63 ± 0.03
0.50 ± 0.06

0.61 ± 0.05
0.56 ± 0.04
0.66 ± 0.04

0.69 ± 0.05
0.71 ± 0.10
0.60 ± 0.05

1.95 ± 0.08
2.02 ± 0.08
2.84 ± 0.04

prised of Sites 4, 5, 9, 10, 12, 14 and 16 showing the water

0.53 ± .14

quality information on deep areas moderately polluted by

DTN

nitrogen. Cluster 3 consisting of Sites 1–3 and 6 showed the

water quality in shallow and relatively isolated areas with
heavy pollution brought about by organic matter, nitrogen
35 ± 22 0.98 ± 0.07
21 ± 13 0.89 ± 0.08
21 ± 10 1.07 ± 0.13

1.17 ± 0.33

1.07 ± 0.07

1.15 ± 0.09
1.07 ± 0.04
1.10 ± 0.09

20 ± 12 2.22 ± 0.05
2.31 ± 0.08
29 ± 14 3.08 ± 0.09
1.08 ± .09

and phosphorus. In this season, the water quality was mainly

determined by the biophysical and chemical processes
TN

inside the lake rather than land-based pollution sources.

32 ± 5
34 ± 5
34 ± 5

27 ± 5
21 ± 7
25 ± 8

15 ± 5
COD

3.3. Pollutant source identification

Kaiser-Meyer-Olkin (KMO) and Bartlett’s Sphericity
325 ± 1 41 ± 11
329 ± 2 26 ± 3
326 ± 3 29 ± 4

240 ± 1 17 ± 4

288 ± 3 23 ± 5
288 ± 0 17 ± 1
291 ± 3 23 ± 7

333 ± 1 11 ± 3

test were used to check the suitability of data for PCA

239 ± 2 9 ± 3
238 ± 1 7 ± 4

335 ± 2 4 ± 3
353 ± 2 6 ± 2

prior to analysis. As KMO approaches 1, PCA of vari-

SS

ables is meaningful; whereas, as it approaches 0, PCA is

unsuitable. In the present study, KMO is 0.635. Bartlett’s
EC

Sphericity test was used to figure out whether the vari-

ables are uncorrelated. The significance level was lower
7.09 ± 0.48
9.02 ± 0.40
9.07 ± 0.13

5.80 ± 0.46
6.51 ± 0.60

6.66 ± 0.52
6.43 ± 0.19
7.34 ± 0.93

6.60 ± 0.14
7.19 ± 0.20
6.87 ± 0.28

than 0.05, indicating that PCA was suitable for variables.

6.45 ± .46

In general, PC loadings are categorized as ‘strong’, ‘mod-

erate’ and ‘weak’ corresponding to absolute values of >0.75,
DO

0.75–0.50 and 0.50–0.30, respectively [39]. As shown in

Table 5, PCA on four seasons data sets have four PCs for
7.70 ± 0.07

8.46 ± 0.05

7.70 ± 0.13
7.74 ± 0.08
7.43 ± 0.16

8.19 ± 0.08
8.40 ± 0.01
7.73 ± 0.34

7.50 ± 0.09
7.36 ± 0.15
7.43 ± 0.16
7.97 ± .32

spring and summer, five PCs for autumn and three for win-
ter with eigenvalues >1, explaining 76.83%, 71.08%, 76.57%
pH

and 80.02% of the total variance in each respective data set,

respectively.
24.8 ± 0.8
24.4 ± 0.5
24.9 ± 0.1

32.4 ± 0.0
32.6 ± 0.2
33.4 ± 0.4

20.3 ± 0.1
20.6 ± 0.0
20.2 ± 0.1

For the spring data set, among four VFs, VF1 accounts
9.0 ± 0.2
9.1 ± 0.2
9.1 ± 0.2

for 31.64% of the total variance, with strong positive load-

WT

ing on pH, SS and TP, and moderate positive loading on

DO and TN, mainly representing SS and nutrient pollu-
tion. The growth of algae in spring, accompanied by the
Cluster 1
Cluster 2
Cluster 3

Cluster 1
Cluster 2
Cluster 3

Cluster 1
Cluster 2
Cluster 3

Cluster 1
Cluster 2
Cluster 3
Summer

Autumn

changes of pH, DO, etc. enhancing the release of nutrients

Winter
Spring
Table 4

from sediments. In addition, VF2 has moderate positive

loading on COD, DTN and PO4–P, which were possibly
 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149 145

Fig. 5. Dendrogram showing clustering of sampling sites from Yushan Lake.

leached from the degradation of plant debris, previously transformation of nitrogen. During the sampling trip,
washed into the lake and accumulated on/in sediments. NO3–N was negatively related to pH while positively
VF3 is moderately close to EC, which was related to the with DO, which means that the nitrification might be the
mineral component of lake water quality. important process occurring in the lake. VF3 explaining
For the data set of summer, VF1, VF2, VF3 and VF4 11.89% of the total variance, has strong loading on SS, COD,
explain 31.27%, 14.64%, 13.50% and 11.67% of the total vari- together with VF4 (10.92% of the total variance) governed
ance, respectively. VF1 has strong positive loading with WT, by COD and TN, attributed to the entrance of plant litter in
SS and TP, and VF2 has strong positive loading on NO3–N. autumn.
VF1 and VF2 are considered anthropogenic pollution sources Finally, for the data set related to winter, VF1 explaining
because the non-point source pollution got more serious due 50% of the total variance, has significantly strong loading
to the frequent precipitation. VF3 has moderated positive on EC, TN, DTN, NH4–N, NO3–N and PO4–P, and moderate
loading on pH and COD, and VF4 on COD and TN, mainly loading on TP and DTP. VF2 occupying 15.90% of the total
representing organic matter pollution. The input of nutri- variance, is moderately related to pH and SS. Whereas VF3
ents enhanced the growth of phytoplankton, which had high has strong loading on WT and moderate loading on SS. VF2
biomass in this season. Even though the photosynthesis was and VF3 can explain the mineralization of organic matter
strong, DO centration was not high, due to the depletion by using DO, producing suspended solids. Also, the rare pre-
increased mineralization of organic matter [40]. cipitation coupled with continuous evaporation resulted
As to autumn, the data set was figured out five VFs. in a considerable decrease in water depth. Moreover, the
VF1 (31.14% of the total variance) having loading on WT, biogeochemical process in the lake was affected by the tem-
pH and NO3–N, representing the internal migration and perature. Consequently, the phenomenon of high dissolved
146 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

nutrient levels together with low pH and high EC took result, the measurement improving water quality is provided
place. in Fig. 6.
Sites 11–14 and 16 are situated on the west edge of a
3.4. Implication for lake water quality management park. Several ditches go across the part introducing land-
based plant litter with stormwater into the lake. During both
Based on the information available from CA and PCA, in spring and summer, SS and TP had significant relation-
lake water quality impairment was mainly related to SS, ship (r = 0.824, p < 0.001 for spring; r = 0.767, p = 0.001 for
phosphorus and nitrogen pollution, which mainly arose in summer). Therefore, it is suggested to construct ecologi-
those places where stormwater was discharged or there was cal engineering in the ditches, eliminating the entrance of
a stagnant area with weak water exchange. In detail, lake phosphorus carried out by stormwater generated from the
water quality was governed by SS and TP at the sites (11– park. Also, there is a discharge outlet near Site 15 for storm-
16) grouped into Cluster 3 in spring and summer, NO3–N water from the urban impervious catchment. The input
at the sites (2–4) from Cluster 2 in autumn, while nitrogen of pollutants can be contained via using post low impact
and phosphorus of Cluster 2 (Sites 1–3 and 6) in winter. As a development (LID) strategy. Sites 1–6 mainly represents

Table 5
Loading of variables on significant principal components for seasonal data sets

Variable VF1 VF2 VF3 VF4 VF5

Spring (four significant components)
WT –0.116 0.197 0.247 –0.572
pH 0.808 –0.094 0.027 –0.032
DO 0.535 –0.383 –0.519 0.125
EC –0.561 –0.107 0.603 0.167
SS 0.825 0.065 0.031 0.209
COD –0.349 0.585 –0.282 –0.542
TN 0.582 0.459 0.482 0.345
DTN –0.196 0.713 0.452 0.399
NH4–N –0.378 –0.617 –0.167 0.406
NO3–N –0.817 –0.027 0.149 0.274
TP 0.878 0.125 0.222 0.004
DTP –0.293 0.392 –0.628 0.436
PO4–P 0.022 0.733 –0.6020 0.167
Eigenvalue 4.113 2.361 2.069 1.446
% Total variance 31.64 18.16 15.91 11.12
Cumulative % Total variance 31.64 49.80 65.71 76.83
Summer (four significant components)
WT 0.830 –0.148 0.018 0.227
pH –0.588 –0.411 0.539 –0.030
DO 0.269 –0.331 –0.622 0.150
EC 0.354 0.154 0.357 –0.404
SS 0.818 0.389 0.079 –0.266
COD 0.148 –0.356 0.543 0.542
TN –0.290 0.233 –0.458 0.604
DTN 0.420 0.469 0.271 0.475
NH4–N 0.454 –0.005 0.442 0.403
NO3–N –0.231 0.840 0.220 –0.158
TP 0.897 0.185 –0.187 0.118
DTP –0.542 0.453 –0.215 0.232
PO4–P –0.723 0.281 0.202 0.263
Eigenvalue 4.065 1.903 1.755 1.517
% Total variance 31.27 14.64 13.50 11.67
Cumulative % Total variance 31.27 45.91 59.41 71.08

(Continued)
 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149 147

Table 5 Continued

Variable VF1 VF2 VF3 VF4 VF5

Autumn (five significant components)
WT 0.827 0.048 –0.249 –0.108 0.238
pH 0.800 0.035 0.300 0.024 –0.442
DO –0.587 –0.468 –0.362 0.147 –0.020
EC –0.356 0.379 0.198 –0.466 0.590
SS –0.378 0.154 0.536 –0.041 –0.496
COD –0.283 0.247 0.606 0.577 0.206
TN –0.330 –0.237 0.152 0.577 0.451
DTN 0.406 –0.529 0.017 0.042 0.298
NH4–N 0.620 –0.505 0.073 0.341 –0.083
NO3–N 0.785 0.208 0.144 0.256 0.109
TP –0.655 –0.222 –0.321 0.163 –0.284
DTP –0.192 0.653 –0.395 0.480 –0.121
PO4–P 0.400 0.465 –0.510 0.245 0.027
Eigenvalue 3.918 1.777 1.545 1.419 1.295
% Total variance 30.14 13.67 11.89 10.91 9.96
Cumulative % Total variance 30.14 43.81 55.70 66.61 76.57
Winter (five significant components)
WT 0.016 –0.433 0.760
pH –0.031 0.511 –0.713
DO –0.042 –0.935 –0.202
EC 0.981 –0.020 –0.043
SS –0.101 0.688 0.630
COD 0.483 0.237 –0.155
TN 0.981 –0.034 –0.018
DTN 0.980 –0.038 –0.090
NH4–N 0.957 0.082 –0.077
NO3–N 0.961 –0.089 –0.058
TP 0.619 0.395 0.359
DTP 0.740 –0.181 0.213
PO4–P 0.833 –0.091 –0.007
Eigenvalue 6.596 2.066 1.741
% Total variance 50.74 15.89 13.39
Cumulative % Total variance 50.74 66.63 80.02

lake water quality of stagnant area during dry the season. • The water quality of Yushan Lake had a significant sea-
Hence, water exchange inside the lake between the stagnant sonal variation with the heaviest contamination in winter.
area and the deep-water area should be prompted like opti- Overall, the water quality was class IV in spring, summer
mizing the canal or other measurements. On the other hand, and autumn, while poor V in winter.
ecological engineering also can be conducted in the stagnant • CA categorized sampling sites into three clusters regard-
area having Sites 1–6. In addition, the washing activity of less of the season. In spring the release inside the lake is
residents is taken into the management. believed to be the major factor governing the water qual-
ity. Due to the lack of stormwater control, the entrance
4. Conclusions of stormwater from different land-use types affected the
water quality deeply over summer. During autumn, leaf
With the increasing concern on surface water quality, litter of trees went into the lake. As it was subjected to
there is a tremendous need to disclose details of water qual- soaking, the materials, like nutrients, from plants were
ity appropriately. In the present study, PCA and CA were leached into the water. As to winter, attribute to the rare
integrated to investigate the water quality of Yushan Lake. precipitation and continuous evaporation, the water
The major conclusions are as follows: volume and depth decreased greatly. Moreover, the
148 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149

Fig. 6. Strategy improving lake water quality based on multivariate statistical approaches.

biogeochemical process in the lake was suppressed by References

the low temperature. As a result, a high level of dissolved [1] H.T. Aboelnga, L. Ribbe, F.-B. Frechen, J. Saghir, Urban water
nutrients was observed. security: definition and assessment framework, Resources,
• Based on the result of multivariate statistical approaches, 8 (2019) 178, doi: 10.3390/resources8040178.
the strategy to improve the water quality was suggested. [2] U. Schirpke, L. Egarter Vigl, E. Tasser, U. Tappeiner, Analyzing
Construct ecological engineering in the ditches link with spatial congruencies and mismatches between supply, demand
and flow of ecosystem services and sustainable development,
the lake to reduce the input of phosphorus. The post- Sustainability, 11 (2019) 2227, doi: 10.3390/su11082227.
low impact development strategy should to the taken to [3] Q. Ding, L. Wang, M. Fu, N. Huang, An integrated system
contrail the stormwater generated from the impervious for rapid assessment of ecological quality based on remote
areas. Inside the lake, ecological engineering in the stag- sensing data, Environ. Sci. Pollut. Res., 27 (2020) 32779–32795.
nant area and the circulation of water also are needed. [4] X. Wang, Y. Chen, Z. Li, G. Fang, Y. Wang, Development and
utilization of water resources and assessment of water security
Also, the plant litter, particularly that from the park, and in Central Asia, Agric. Water Manage., 240 (2020) 106297,
the washing of residents should be controlled effectively. doi: 10.1016/j.agwat.2020.106297.
[5] R. Hu, Pollution control and remediation of rural water
Acknowledgments resource based on urbanization perspective, Environ. Technol.
Innovation, 20 (2020) 101136, doi: 10.1016/j.eti.2020.101136.
This research was supported by the open fund by Jiangsu [6] H. Yasmeen, Q. Tan, H. Zameer, J. Tan, K. Nawaz, Exploring
Key Laboratory of Atmospheric Environment Monitoring the impact of technological innovation, environmental
and Pollution Control, A Project Funded by the Priority regulations and urbanization on ecological efficiency of China
in the context of COP21, J. Environ. Manage., 274 (2020) 111210,
Academic Program Development of Jiangsu Higher Education doi: 10.1016/j.jenvman.2020.111210.
Institutions (PAPD), the Natural Science Foundation of Anhui [7] D. Kirikkaleli, J.K. Sowah, A wavelet coherence analysis:
Province (Grant No. 1808085QD109 and 1908085QD167) and nexus between urbanization and environmental sustainability,
the National Natural Science Foundation of China (Grant Environ. Sci. Pollut. Res., 27 (2020) 30295–30305.
No. 41701553), the Scientific Research Foundation for the [8] A.R. Afitiri, F.A. Armah, E.K. Afrifa, B. Ekumah, The joint
effect of natural and human-induced environmental factors
Returned Overseas Chinese Scholars of Anhui Province. on surface water quality in the Birim North District of Ghana,
The authors are grateful for the support. Water Pract. Technol., 15 (2020) 605–618.
[9] A. Mebarkia, A. Boufekane, Human activity impact on surface
water quality in semi-arid regions: a case study of Aïnzeda lake
Declaration of interest statement (North-East Algeria), Water Supply, 20 (2020) 1726–1744.
[10] M. Varol, Spatio-temporal changes in surface water quality
The authors declare that they have no known compet-
and sediment phosphorus content of a large reservoir in
ing financial interests or personal relationships that have Turkey, Environ. Pollut., 259 (2020) 113860, doi: 10.1016/j.
appeared to influence the work reported in this paper. envpol.2019.113860.
 S. Niu et al. / Desalination and Water Treatment 239 (2021) 137–149 149

[11] C. Paijens, A. Bressy, B. Frère, B. Tedoldi, D. Mailler, R. Rocher, study in Laoheba phosphorite mine in Sichuan, China, Arabian
V.P. Neveu, R. Moilleron, Urban pathways of biocides towards J. Geosci., 7 (2014) 3973–3982.
surface waters during dry and wet weathers: assessment at the [26] P. Li, R. Tian, R. Liu, Solute geochemistry and multivariate
Paris conurbation scale, J. Hazard Mater., 402 (2020) 123765, analysis of water quality in the Guohua Phosphorite Mine,
doi: 10.1016/j.jhazmat.2020.123765. Guizhou Province, China, Exposure Health, 11 (2019) 81–94.
[12] Y. Li, S. Cui, B. Gao, J. Tang, W. Huang, Y. Huang, Modeling [27] J. Wu, P. Li, D. Wang, X. Ren, M. Wei, Statistical and multivariate
nitrogen flow in a coastal city—a case study of Xiamen in statistical techniques to trace the sources and affecting factors
2015, Sci. Total Environ., 735 (2020) 139294, doi: 10.1016/j. of groundwater pollution in a rapidly growing city on the
scitotenv.2020.139294. Chinese Loess Plateau, Hum. Ecol. Risk Assess., 26 (2020)
[13] N. Aryal, J. Wood, I. Rijal, D. Deng, M.K. Jha, A. Ofori-Boadu, 1603–1621.
Fate of environmental pollutants: a review, Water Environ. [28] Y. Anteneh, G. Zeleke, E. Gebremariam, Assessment of surface
Res., 92 (2020) 1587–1594. water quality in legendaria and Dire catchments, Central
[14] P. Li, W. Feng, C. Xue, R. Tian, S. Wang, Spatiotemporal Ethiopia, using multivariate statistical analysis, Acta Ecol. Sin.,
variability of contaminants in lake water and their risks to 38 (2018) 81–95.
human health: a case study of the Shahu Lake tourist area, [29] A.D. Yetis, F. Akyuz, Water quality evaluation by using
northwest China, Exposure Health, 9 (2017) 213–225. multivariate statistical techniques and pressure‑impact
[15] Q. Quan, R. Tong, W. Sun, Y. Zhao, J. Yu, G. Wang, S. Shrestha, analysis in wetlands: Ahlat Marshes, Turkey, Environ. Dev.
Y. Jin, Anthropogenic influences on the water quality of the Sustainability, 23 (2021) 969–988.
Baiyangdian Lake in North China over the last decade, Sci. Total [30] G. Qin, J. Liu, S. Xu, T. Wang, Water quality assessment and
Environ., 701 (2020) 134929, doi: 10.1016/j.scitotenv.2019.134929. pollution source apportionment in a highly regulated river
[16] J. Wu, C. Xue, R. Tian, S. Wang, Lake water quality assessment: of Northeast China, Environ. Monit. Assess., 192 (2020) 446,
a case study of Shahu Lake in the semi-arid loess area of doi: 10.1007/s10661-020-08404-0.
northwest China, Environ. Earth Sci., 76 (2017) 232, doi: 10.1007/ [31] T. Marara, L.G. Palamuleni, A spatiotemporal analysis of
s12665-017-6516-x. water quality characteristics in the Klip river catchment, South
[17] J. Wang, Z. Fu, H. Qiao, F. Li, Assessment of eutrophication and Africa, Environ. Monit. Assess., 192 (2020) 578, doi: 10.1007/
water quality in the estuarine area of Lake Wuli, Lake Taihu, s10661-020-08441-9.
China, Sci. Total Environ., 650 (2019) 1392–1402. [32] A. Barakat, M. El Baghdadi, J. Rais, B. Aghezzaf, M. Slassi,
[18] A. Davraz, S. Varol, E. Sener, S. Sener, F. Aksever, B. Kirkan, Assessment of spatial and seasonal water quality variation of
A. Tokgözlü, Assessment of water quality and hydrogeochemical Oum Er Rbia River (Morocco) using multivariate statistical
processes of Salda alkaline lake (Burdur, Turkey), Environ. techniques, Int. Soil Water Conserv. Res., 4 (2016) 284–292.
Monit. Assess., 191 (2019) 71, doi: 10.1007/s10661-019-7889-y. [33] F. Ustaoğlu, Y. Tepe, Water quality and sediment contamination
[19] A.U. Alam, D. Clyne, H. Jin, N. Hu, M.J. Deen, Fully integrated, assessment of Pazarsuyu Stream, Turkey using multivariate
simple, and low-cost electrochemical sensor array for in situ statistical methods and pollution indicators, Int. Soil Water
water quality monitoring, ACS Sens., 5 (2020) 412–422. Conserv. Res., 7 (2018) 47–56.
[20] E. Kumpel, C. MacLeod, K. Stuart, A. Cock-Esteb, P. Khush, [34] M. Sebastián Gradilla-Hernández, J. de Anda, A. Garcia-
R. Peletz, From data to decisions: understanding information Gonzalez, D. Meza-Rodríguez, C. Yebra Montes, Y. Perfecto-
flows within regulatory water quality monitoring programs, Avalos, Multivariate water quality analysis of Lake Cajititlán,
npj Clean Water, 3 (2020) 38, doi: 10.1038/s41545-020-00084-0. Mexico, Environ. Monit. Assess., 192 (2020) 5, doi: 10.1007/
[21] K. Haldar, K. Kujawa-Roeleveld, P. Dey, S. Bosu, D.K. Datta, s10661-019-7972-4.
H.H. Rijnaarts, Spatio-temporal variations in chemical-physical [35] I.G. Hatvani, V.D. Barros, P. Tanos, J. Kovács, I.S. Kovács,
water quality parameters influencing water reuse for irrigated A. Clement, Spatiotemporal changes and drivers of trophic
agriculture in tropical urbanized deltas, Sci. Total Environ., 708 status over three decades in the largest shallow lake in
(2020) 134559, doi: 10.1016/j.scitotenv.2019.134559. Central Europe, Lake Balaton, Ecol. Eng., 151 (2020) 105861,
[22] S. Jehan, I. Ullah, S. Khan, S. Muhammad, S.A. Khattak, doi: 10.1016/j.ecoleng.2020.105861.
T. Khan, Evaluation of the Swat River, Northern Pakistan, [36] X. Hou, L. Feng, J. Tang, X.P. Song, B.A. Bryan, Anthropogenic
water quality using multivariate statistical techniques and transformation of Yangtze plain freshwater lakes: patterns,
water quality index (WQI) model, Environ. Sci. Pollut. Res., drivers and impacts, Remote Sens. Environ., 248 (2020) 111998,
27 (2020) 38545–38558. doi: 10.1016/j.rse.2020.111998.
[23] S.M. Njuguna, J.A. Onyango, K.B. Githaiga, R.W. Gituru, [37] M. Kumarin, S. Tripathin, V. Pathakn, B.D. Tripathin,
X. Yan, Application of multivariate statistical analysis and Chemometric characterization of river water quality, Environ.
water quality index in health risk assessment by domestic use Monit. Assess., 185 (2013) 3081–3089.
of river water. Case study of Tana River in Kenya, Process Saf. [38] H. Yang, Z. Shen, J. Zhang, W. Wang, Water quality
Environ. Prot., 133 (2020) 149–158. characteristics along the course of the Huangpu River, J. Integr.
[24] B.A. Berhe, Evaluation of groundwater and surface water Environ. Sci., 19 (2007) 1193–1198.
quality suitability for drinking and agricultural purposes in [39] C.W. Liu, K.H. Lin., Y.M. Kuo, Application of factor analysis in
Kombolcha town area, eastern Amhara region, Ethiopia, Appl. the assessment of groundwater quality in a blackfoot disease
Water Sci., 10 (2020) 127, doi: 10.1007/s13201-020-01210-6. area in Taiwan, Sci. Environ., 313 (2003) 77–89.
[25] J. Wu, P. Li, H. Qian, Z. Duan, X. Zhang, Using correlation and [40] X. Jiang, S. Wang, L. Zhong, X. Jin, S. Sun, Seasonal variation
multivariate statistical analysis to identify hydrogeochemical characteristics of Algae biomass in Chaohu Lake, Environ. Sci.,
processes affecting the major ion chemistry of waters: case 31 (2010) 2056–2062 (in Chinese).