Water Quality Dynamics in the Zhuxihe River Basin in Hainan Province, China: Insights from Temporal and Spatial Analysis

Abstract

The Zhuxihe River has faced significant water quality challenges in recent years. Although control measures have been implemented, the pollution levels remain concerning. This paper aims to investigate the spatio-temporal variations in the water quality of the Zhuxihe River through field sampling, chemical testing, and synthetic evaluation. We collected 52 water samples in both dry and wet seasons along the main river and its tributaries. The evaluation, which utilized the integrated SFE-FCE method, identified MnO₄²⁻, NH₃-N, TP, and TFe as the primary pollutants. In the dry season, MnO₄²⁻ concentrations ranged from 1.6 mg/L to 19.8 mg/L, NH₃-N ranged from 0.12 mg/L to 2.04 mg/L, and TP concentrations varied from 0.1 mg/L to 5.61 mg/L. In the wet season, MnO₄²⁻ concentrations ranged from 4.9 mg/L to 13.9 mg/L, NH₃-N from 0.27 mg/L to 1.73 mg/L, and TP from 0.07 mg/L to 1.31 mg/L. The results indicate the concentrations of MnO₄²⁻ are higher in the wet season, and TP and NH₃-N show seasonal fluctuations. Spatially, the downstream section of the main river faces the highest pollution levels. This study provides insights into the pollution dynamics of the Zhuxihe River, offering a scientific foundation for future research and water quality management strategies.

1. Introduction

Rivers have been central to human civilization, serving as cradles for early societies and sustaining various aspects of life [1,2,3,4,5]. However, rapid industrialization and urbanization have resulted in significant pollution of many of these vital waterways, posing challenges to environmental sustainability and public health [6,7,8]. In response to the escalating water pollution crisis, the Ministry of Ecology and Environment of the People’s Republic of China (MEE) issued the ‘Environmental Quality Standards for Surface Water (GB3838-2002)’, which establishes a five-tier classification framework for surface water quality [9]. Grade I represents the best quality and is suitable for source water and national nature reserves. Grade II is of good quality and is applicable to centralized domestic drinking water sources and habitats for rare aquatic species. Grade III is of fairly good quality and is designated for secondary protection zones for drinking water and recreational areas involving direct human contact. Grades IV and V correspond to water suitable for general industrial use and agricultural purposes, respectively, with grade IV representing the poor quality and grade V representing the worst quality.

Hainan Island (Hainan Province), the second-largest island in China, is endowed with valuable freshwater resources. However, several rivers have encountered varying degrees of pollution in Hainan Province [10,11,12,13]. Among these rivers, the Zhuxihe River has faced severe water quality issues in recent years. Although the local government publishes the ‘Hainan Ecological and Environmental Bulletin’ quarterly and annually, which identifies key chemical factors affecting the water quality of the Zhuxihe River, there remains uncertainty regarding temporal and spatial variations of water quality and the primary sources of pollution. This ambiguity has impeded effective remediation efforts to improve the water quality of the Zhuxihe River.

Previous studies have employed various methods to assess water quality, such as the Single Factor Evaluation (SFE) method, the Fuzzy Comprehensive Evaluation (FCE), and artificial intelligence-based (AI-based) methods. Among these, the SFE method is a widely used approach that evaluates water quality based on individual pollutant parameters [14]. By comparing measured concentrations with established water quality standards, this method identifies the most critical pollutant affecting water quality [15,16]. However, the exclusive use of the SFE method presents limitations, as it does not consider the cumulative effects of multiple pollutants and their interactions [17]. Additionally, it is overly sensitive to the worst-performing parameter, which may lead to an oversimplified assessment of overall water quality [14]. To overcome these limitations, Ji [18] conducted a comparative analysis of seven water quality assessment methods to evaluate the impairment of water quality in the Wen-Rui Tang River, located in eastern China. Their findings concluded that the integration of the SFE method with the FCE method effectively addressed uncertainties and yielded a more nuanced evaluation. In addition, AI-based methods, such as machine learning and deep learning methods, are also widely used in the study of water quality [19]. For example, Nasir [20] employed stacking machine learning models to analyze 1679 water quality samples collected over a nine-year period. Their findings indicate that machine learning models can assist researchers in their efforts to improve water quality. Etheram [21] utilized deep learning models to classify water quality, employing a dataset collected from multiple monitoring stations. Their study demonstrates that both XGBoost and LSTM models significantly outperformed other approaches. Although these AI-based methods are promising, they require extensive datasets and substantial computational resources, which are currently unavailable for the Zhuxihe River study.

Despite the significant pollution impacting the water quality of the Zhuxihe River, there has been an absence of systematic studies focusing on quantitative analyses of chemical pollutants. The primary aim of this paper is to investigate the temporal and spatial variations in the surface water quality (SWQ) and the centralized domestic drinking water quality (DWQ) of the Zhuxihe River through sampling, testing, and synthetic evaluation. The sampling process was conducted along the main river and its tributaries in April and August 2023, resulting in 52 groups of water samples collected over these two months and subsequently sent for testing. The sampling and testing processes strictly adhered to the requirements of Chinese national standards, yielding accurate data for evaluation. Although more advanced methods could be used for evaluation, their application often requires computational resources, complex parameter selection, and extensive data, such as multi-year seasonal data and samples covering a broader range of pollutants. Due to data limitations and research focus, this study used the SFE-FCE method, while more advanced approaches could further refine water quality evaluations.

2. Study Area

2.1. Location and the River Basin in the Study Area

Hainan Province, situated in southern China, lies between latitudes 18°10′–20°10′ N and longitudes 108°37′–111°03′ E (see Figure 1a,b). It is the country’s smallest and southernmost province, comprising Hainan Island and its surrounding islets. The island’s extensive coastline and abundant water resources, including rivers such as the Zhuxihe River, are crucial to the region’s socio-economic development.

The Zhuxihe River, the second largest river in Wenchang City, Hainan Province, is one of the few rivers in the region that flows directly into the sea. As illustrated in Figure 1c, the river flows from southeast to northwest, passing through four towns and one village, including Wengtian Town, Fengpo Town, Jinshan Town, Puqian Town, and Hushan Village. The main river is divided into three sections: the upstream section in Fengpo Town, the midstream section in Jinshan Town, and the downstream section in Puqian Town. The total basin area is 392 km², encompassing four primary tributaries: Paigang Stream, Baimang Stream, Shangxi Ditch, and Jiaolong Ditch.

As shown in Figure 1c, a total of 26 sites along the main river and its tributaries were selected for water sampling, including 20 sites along the main river and 6 sites along the tributaries. The upstream section of the main river includes sites SW01, SW02, SW03, SW05, SW06, SW07, and SW09. The midstream section comprises sites SW10, SW12, SW13, SW14, and SW15. The downstream section contains sites SW16, SW17, SW18, SW19, SW20, SW22, SW23, and SW24. Additionally, the tributary sites include site SW21, located in the Paigangxi Stream; sites SW11, SW25, and SW26, located in the Baimangxi Stream; site SW8, located in the Shangxigou Ditch; and site SW4, located in the Jiaolonggou Ditch.

2.2. Climate and Seasonal Water Sample Collection

Wenchang City, where Zhuxihe River Basin is located, is characterized by a tropical monsoon climate that features warm and humid conditions throughout the year, accompanied by abundant precipitation. The average annual precipitation in the Zhuxihe River Basin ranges between 1500 and 2100 mm. Surface water in the Basin is predominantly recharged by rainfall. Rainfall is primarily concentrated during the summer and autumn months, specifically from May to October, while significantly less precipitation occurs during the winter months, from November to April of the following year. To investigate the water quality of the Zhuxihe River during wet and dry periods, April and August were selected as representative months for water sampling in the dry and wet seasons, respectively.

2.3. General Condition and Possible Causes of Deterioration of Water Quality

According to the ‘Hainan Ecological and Environmental Bulletin, 2023’, published by the Department of Ecology and Environment of Hainan Province (HEE), the Zhuxihe River has been experiencing moderate to severe pollution, primarily characterized by elevated levels of the permanganate index and total phosphorus [22].

The causes of pollution in Zhuxi River are complicated, and a unified conclusion has not yet been reached so far. Possible causes contributing to the deterioration of water quality can be summarized as follows:

(1)

Small-scale and non-industrial livestock and aquaculture

Since the 1980s, livestock and aquaculture farming along the Zhuxihe River has developed, establishing Wenchang City as one of the primary agricultural bases in Hainan Province [23]. A non-public statistical study conducted by the local government revealed that most livestock and aquaculture farms are small-scale, non-industrial operations. This situation leads to the direct discharge of significant quantities of livestock manure and aquaculture wastewater into the river, thereby degrading the water quality and the eco-environment of the Zhuxihe River.

(2)

Traditional farming methods

Agriculture in the Zhuxihe River Basin primarily relies on traditional farming methods. For example, fertilization is mainly conducted through soaking and broadcasting, while irrigation predominantly employs the flood-based approach. Consequently, irrigation runoff with unabsorbed pesticides and fertilizers is directly discharged into the river, causing excessive concentrations of nitrogen and phosphorus in the river.

(3)

Natural geographical characteristics

The main river extends 46.17 km with an average channel slope of only 0.32‰, resulting in a flow velocity of less than 0.5 m/s throughout the year. The four tributaries possess drainage areas of less than 100 km², contributing to limited water inflows. This combination of low flow velocity and limited catchment areas results in poor hydrodynamic conditions and prolonged water renewal cycles. Consequently, the dispersion and dilution of pollutants are hindered, leading to their tendency to settle and adsorb onto the riverbed sediments, gradually accumulating over time.

(4)

Ineffective mitigation measures

Since 2018, Wenchang City has implemented various measures to mitigate pollution. These measures included regulating pollution sources from domestic sewage, aquaculture, and livestock farming, as well as constructing wastewater treatment plants and initiating pollution remediation projects. However, insufficient funding and inadequate maintenance have impeded the effectiveness of these measures, resulting in limited improvements in water quality.

3. Materials and Methods

3.1. Water Sample Collection and Testing

The process of water sample collection and testing consisted of three steps: sample collection, sample storage, and sample testing.

In both April and August, we collected 26 groups of water samples from 26 sites, with each site yielding one water sample. This resulted in a total of 52 groups of surface water samples collected from these sites over the two months. Collected samples were then categorized into volatile organic compounds (VOCs), semi-volatile organic compounds (SVOCs), stable organics, heavy metals, and inorganics during bottling. Disposable droppers added specific reagents, and all bottles were sealed and labelled for traceability.

Immediately after collection, samples were stored in the incubator containing ice packs. Upon arrival at the laboratory within 12 h, the samples were transferred to a refrigerated cabinet maintained at 4 °C, ensuring conditions aligned with the specified storage requirement [24]. All samples were tested within their designated storage periods to prevent deterioration or contamination.

Tests were designed to minimize error and ensure data reliability. For each group of analyses, two blank tests were conducted to control potential reagent-related errors. Additionally, reference materials were added to one to two samples as internal benchmarks for accuracy validation. A further 20% of the total samples were randomly selected for repeated analysis to ensure quality control. In case of abnormal results, these tests were repeated.

3.2. Examination of Data Reliability

The data obtained from the tests required further examinations of anion and cation balance, as well as total hardness balance, to ensure compliance with the required criteria before proceeding with subsequent analyses.

(1)

Anion and cation balance examination

Natural water, when utilized as an electrolyte solution, exhibits charge neutralization. Therefore, the following charge neutralization equations must be satisfied in the collected water samples:

$${\displaystyle \sum Z{m}_c={\displaystyle \sum Z{m}_a}}$$

(1)

where m_c and m_a represent the molar concentrations of cations and anions, respectively, mmol/L; Z is the charge number of the ions. Due to errors that may occur during the testing and analyzing, when both ∑Zm_c and ∑Zm_a in the water solution exceed 5 mmol/L, a relative error in charge balance of less than 5% is required to ensure data reliability, shown as follows:

$$\left|\mathrm{E}\right|={\displaystyle \frac{\left|{\displaystyle \sum Z{m}_c-{\displaystyle \sum Z{m}_a}}\right|}{\left({\displaystyle \sum Z{m}_c+{\displaystyle \sum Z{m}_a}}\right)}}\times 100\%\le 5\%$$

(2)

(2)

Total hardness examination

The total hardness in natural water is primarily determined by the concentrations of Ca²⁺ and Mg²⁺. It is generally expressed as the equivalent concentration of CaCO₃ and can be calculated using the following equation:

$$T=D\times 50$$

(3)

where D represents the sum of the milligram equivalents of Ca²⁺ and Mg²⁺, meq/L, and T is the total hardness (mg/L, expressed as CaCO₃).

To ensure the reliability of the measurement data, the absolute value of the relative error |E| between the calculated total hardness and the measured Ca²⁺ and Mg²⁺ concentrations should be less than 1%, shown as follows:

$$\left|\mathrm{E}\right|={\displaystyle \frac{\left|{\displaystyle \sum D\times 50-{\displaystyle \sum T}}\right|}{\left({\displaystyle \sum D\times 50+{\displaystyle \sum T}}\right)}}\times 100\%\le 1\%$$

(4)

3.3. Water Quality Evaluation System

3.3.1. Evaluation Factors

A total of 52 groups of water samples were collected from sampling sites along the Zhuxihe River and sent to the Hainan Provincial Geological Testing and Research Center for testing. The selection of the tested chemical factors was guided by the ‘Environmental Quality Standards for Surface Water (GB3838-2002)’, which provides a comprehensive framework for evaluating water quality [9].

As shown in

Table 1. Evaluation factors of water quality of Zhuxihe River.

Types of Factors	Factors	Numbers
SWQ basic factors	MnO42−, NH3-N, TP, Cu, Zn, F−, Se, As, Hg, Cd, Cr, Pd, CN−, VPs, S2−	15
DWQ supplementary factors	SO42−, Cl−, NO3−, TFe, Mn	5

, evaluation factors include 15 SWQ basic factors and 5 DWQ supplementary factors. The SWQ basic factors are factors such as permanganate index (MnO₄²⁻), ammonia nitrogen (NH₃-N), total phosphorous (TP), copper (Cu), zinc (Zn), fluoride (F⁻), selenium (Se), arsenic (As), mercury (Hg), cadmium (Cd), hexavalent chromium [Cr(VI)], palladium (Pd), cyanide (CN⁻), volatile phenolics (VPs), and sulfide (S²⁻). For DWQ, five additional factors, including sulfate (SO₄²⁻), chloride (Cl⁻), nitrate (NO₃⁻), total iron (TFe), and manganese (Mn), should also be tested to provide a more detailed evaluation of the water quality.

3.3.2. Water Quality Classification

The classification system utilized for assessing the water quality is derived from the five-tier classification framework outlined in the ‘Environmental Quality Standards for Surface Water (GB3838-2002)’ [9]. This system categorizes water quality into five grades (I to V), ranging from the best (grade I) to the worst (grade V). Each grade is determined by specific classification limits of different factors, as detailed in

Table 2. Classification limits of water quality factors.

Factors	Grades
	Best	Good	Fairly Good	Poor	Worst
	I	II	III	IV	V
	Classification Limits (Unit: mg/L)
MnO42−	2	4	6	10	15
NH3-N	0.1	0.5	1	1.5	2
TP	0.02	0.1	0.2	0.3	0.4
Cu	0.01	1	1	1	1
Zn	0.05	1	1	2	2
F−	1	1	1	1.5	1.5
Se	0.01	0.01	0.01	0.02	0.02
As	0.05	0.05	0.05	0.1	0.1
Hg	5 × 10−5	5 × 10−5	1 × 10−4	1 × 10−3	1 × 10−3
Cd	1 × 10−3	5 × 10−3	5 × 10−3	5 × 10−3	0.01
Cr(VI)	0.01	0.05	0.05	0.05	0.1
Pd	0.01	0.01	0.05	0.05	0.1
CN−	5 × 10−3	0.05	0.2	0.2	0.2
VPs	2 × 10−3	2 × 10−3	5 × 10−3	0.01	0.1
S2−	0.05	0.1	0.2	0.5	1
SO42−	50	150	250	350	350
Cl−	50	150	250	350	350
NO3−	2	5	20	30	30
TFe	0.1	0.2	0.3	2	2
Mn	0.05	0.05	0.1	1.5	1.5

. For example, the classification limits for MnO₄²⁻ in grades I, II, III, IV and V are set at 2 mg/L, 4 mg/L, 6 mg/L, 10 mg/L, and 15 mg/L, respectively. Therefore, when the measured concentration of MnO₄²⁻ in a sample is 10 mg/L, the water quality indicated by MnO₄²⁻ is classified as poor.

3.3.3. Synthetic Evaluation

The synthetic evaluation of water quality, containing both SWQ and DWQ, was conducted in two steps. In the first step, the mass concentration of each factor was evaluated using both the SFE method and FCE method to derive the corresponding grade of the water quality. In the second step, the grade of the worst-performing factor among all evaluated factors determined the final synthetic evaluation result of the water quality.

(1)

SFE method

The SFE method relies on the attribute values and their corresponding grades, as outlined in Table 2. Each factor is evaluated on the basis of classification limits, and the boundary range of these limits is utilized to determine the respective water quality category. In cases where the limits for different grades are identical, the evaluation results for the selected factors are assessed according to the principle of Choosing a Superior Alternative (CSA). For example, if the measured concentration of Cu is 1.0 mg/L, the test result corresponds to the attribute values for grades II, III, IV, and V. However, in accordance with the CSA principle, the sample should be categorized as grade II.

The five classification limits for each factor in Table 2 are defined by specific numerical values. However, actual measured data may fall between two grades, leading to potential discrepancies in classification. This subjectivity can result in different conclusions, introducing ambiguity and uncertainty when relying solely on the SFE method. For example, also considering the measured concentration of Cu, the classification limit for grade I is 0.01 mg/L, while grade II has a limit of 1.0 mg/L. In this case, if the measured value is 0.1 mg/L, it presents a challenge in determining its classification, as it is not clear whether it falls under grade I or grade II.

(2)

FCE method

The FCE method establishes membership functions that facilitate fuzzy quantification in the evaluation process. This approach evaluates the degree of subordinates in the measurement value within the evaluation system rather than forcing a classification into a single grade [25,26]. Consequently, it more accurately reflects the position within the quality level, thereby minimizing the errors associated with the sole SFE method and reducing the potential misclassification of quality levels.

A variety of membership functions are available for calculating membership degrees, including rectangular, trapezoidal, parabolic, Gaussian, Cauchy, and ridge functions, which are among the most commonly used [27]. Although individuals may choose different membership functions, employing the same function within a fuzzy evaluation can ensure a consistent representation and yield accurate and reliable outcomes [28,29,30].

Previous studies [14,31,32] have shown that both the trapezoidal and the semi-trapezoidal membership functions are commonly used in elevating water quality. In our study, we used the semi-trapezoidal membership function, which is shown as follows:

$$\begin{array}{l}{\mu }_{i1}\left(x_i\right)=\left\{\begin{array}{cc}1& x_i\le s_{i1}\\ \left(s_{i2}-x_i\right)/\left(s_{i2}-s_{i1}\right)& s_{i1}<x_i<s_{i2}\\ 0& x_i\ge s_{i2}\end{array}\right.\hfill \\ {\mu }_{ij}\left(x_i\right)=\left\{\begin{array}{cc}\left(x_i-s_{ij-1}\right)/\left(s_{ij}-s_{ij-1}\right)& s_{ij-1}<x_i<s_{ij}\\ \left(s_{ij+1}-x_i\right)/\left(s_{ij+1}-s_{ij}\right)& s_{ij}<x_i<s_{ij+1}\\ 0& x_i>s_{ij+1}\end{array}\right.\hfill \\ {\mu }_{im}\left(x_i\right)=\left\{\begin{array}{cc}1& x_i\ge s_{i1}\\ \left(x_i-s_{im-1}\right)/\left(s_{im}-s_{im-1}\right)& s_{im-1}<x_i<s_{im}\\ 0& x_i>s_{im-1}\end{array}\right.\hfill \end{array}$$

(5)

where µ_ij represents the membership degree of the i^th factor to the j^th grade, (i = 1, 2, …, n; j = 1, 2, …, m); n is the number of evaluation factors; m is the number ranging from 1 to 5 that correspond to the grades I to V; x_i is the tested value (actual measured data) of the factors; s_ij is the standard value of the factors corresponding to the grades of I to V.

3.4. Data Analysis and Statistical Tools

ArcGIS 10.8 (Version 10.8, https://www.esri.com/, accessed on 18 January 2025) and MapGIS 6.5 (Version 6.5, https://www.mapgis.com/, accessed on 18 January 2025) were used to prepare figures and perform spatial analysis. Microsoft Excel (VBA Programming) was used for data manipulation and statistical analysis.

4. Results

4.1. Testing Results

(1)

Dry season

Factors such as Cd, Cr, CN⁻, and VPs were not detected in water samples collected during the dry season. In contrast, the remaining 16 factors were detected (see Figure 2). Among these 16 factors, MnO₄²⁻ exhibits significant variation across sites, with a minimum concentration of 1.6 mg/L at site SW24 to a maximum concentration of 19.8 mg/L at site SW25. The average value of concentration for MnO₄²⁻ is 6.44 mg/L. The concentrations of NH₃-N vary from 0.12 mg/L at site SW04 to 2.04 mg/L at site SW25, with an average of 0.66 mg/L. The TP levels also display considerable variability, ranging from 0.1 mg/L at site SW04 to 5.61 mg/L at site SW24, with an average of 0.67 mg/L. Although most heavy metal concentrations are relatively low, some sites exhibit elevated levels. For instance, Hg peaks 1.58 × 10⁻⁴ mg/L at SW25. Inorganic compounds such as SO₄²⁻ and Cl⁻ show extreme values at site SW24, with SO₄²⁻ of 1570 mg/L and Cl⁻ of 1.67 × 10⁴ mg/L. TFe and Mn levels are also elevated at certain sites, including SW25 with TFe of 38.1 mg/L and SW16 with Mn of 0.389 mg/L.

The parameter NSF in

Table 3. Statistical results of tests of water quality factors in the dry season.

Factors	Minimum (mg/L)	Maximum (mg/L)	Average (mg/L)	NSF	RNSF (%)	NOM	RNOM (%)
MnO42−	1.6	19.8	6.44	26	100	8	30.77
NH3-N	0.12	2.04	0.66	26	100	5	19.23
TP	0.1	5.61	0.67	26	100	11	42.31
Cu	6 × 10−4	2.8 × 10−3	1.5 × 10−3	26	100	0	0
Zn	1.5 × 10−3	9.9 × 10−3	4.4 × 10−3	26	100	0	0
F−	0.11	0.84	0.27	26	100	0	0
Se	1.7 × 10−4	4 × 10−4	2.8 × 10−4	18	69.23	0	0
As	4.2 × 10−4	0.0149	2.2 × 10−3	25	96.15	0	0
Hg	8 × 10−5	1.6 × 10−4	1 × 10−4	23	88.46	0	0
Cd	Not detected
Cr	Not detected
Pd	4.6 × 10−4	7.24 × 10−3	2.41 × 10−3	23	88.46	0	0
CN−	Not detected
VPs	Not detected
S2−	2 × 10−3	0.014	5.9 × 10−3	22	84.62	0	0
SO42−	6.2	2100	232	26	100	2	7.70
Cl−	25	16700	1831	26	100	3	11.50
NO3−	0.12	3.9	1.4417	23	88.46	0	0
TFe	0.24	38.1	3.1892	26	100	19	73.08
Mn	0.01	0.389	0.10715	26	100	0	0

represents the number of sample groups containing a specific factor. RNSF in Table 3 is the rate of NSF relative to the total number of sample groups, calculated as RNSF = NSF/26. NOM in Table 3 is defined as the number of sample groups containing a specific factor with a concentration exceeding the classification limit at grade III. RNOM in Table 3 is the rate of NOM relative to the total number of sample groups, calculated as RNOM = NOM/26.

Statistical results in Table 3 indicate that not all sample groups contain the specific factors. As shown, ten factors, including MnO₄²⁻, NH₃-N, TP, Cu, Zn, F⁻, SO₄²⁻, Cl⁻, TFe, and Mn, were detected in all sample groups collected during the dry season, resulting in RNSFs of 100%. Five factors, As, Hg, Pb, NO₃⁻, and S²⁻, were detected in 25, 23, 23, and 23 sample groups, respectively, yielding RNSFs of 96.15%, 88.46%, 88.46%, and 84.62%. Factor Se exhibits the lowest RNSF of 69.23%, with a NSF of 18. Among the 16 identified factors, six factors exhibit measured concentrations that exceeded the classification limits of grade III (see Table 3). They are MnO₄²⁻, NH₃-N, TP, SO₄²⁻, Cl⁻, TFe. NOMs for these factors are 5, 11, 2, 3, and 19, and the corresponding RNOMs are 30.77%, 19.23%, 42.31%, 7.7%, 11.5%, and 73.08%, respectively.

(2)

Wet season

During the wet season, four factors were non-detectable: Se, Cr, CN⁻, and VPs, while the remaining 16 factors were detected across all samples (see Figure 3). Among these 16 factors, MnO₄²⁻ shows high levels at most sites, ranging from 4.9 mg/L at site SW26 to 13.9 mg/L at site SW05, with an average of 10.48 mg/L. The concentration of NH₃-N ranges from 0.27 mg/L at site SW26 to 1.73 mg/L at SW21, with an average of 0.94 mg/L. The concentration of TP varies from 0.07 mg/L at SW26 to 1.31 mg/L at SW25, with an average of 0.33 mg/L. Heavy metals are at lower levels during the wet season, with no factors exceeding the classification limit for grade III. Inorganic compounds, such as SO₄²⁻, Cl⁻, TFe, and Mn, exhibit extreme concentration at sites SW24 and SW16. At site, SW24, the maximum concentration of SO₄²⁻ reaches 437 mg/L, and Cl⁻ reaches 3010 mg/L. At site SW16, the maximum concentration of TFe is 16 mg/L, and Mn is 0.956 mg/L.

As shown in Figure 3 and

Table 4. Statistical results of tests of water quality factors in the wet season.

Factors	Minimum (mg/L)	Maximum (mg/L)	Average (mg/L)	NSF	RNSF (%)	NOM	RNOM (%)
MnO42−	4.9	13.9	10.48	26	100	24	92.31
NH3-N	0.27	1.73	0.94	26	100	16	61.54
TP	0.07	1.31	0.33	26	100	15	57.69
Cu	6.1 × 10−4	2.41 × 10−3	1.41 × 10−3	26	100	0	0
Zn	3.06 × 10−3	0.01622	6.82 × 10−3	26	100	0	0
F−	0.08	0.43	0.19	26	100	0	0
Se	Not detected
As	7 × 10−4	0.0251	4.1 × 10−3	26	100	0	0
Hg	2.5 × 10−5	2.7 × 10−5	2.6 × 10−5	3	11.54	0	0
Cd	6 × 10−5	6 × 10−5	6 × 10−5	1	3.85	0	0
Cr	Not detected
Pd	3.2 × 10−4	5.3 × 10−3	1.7 × 10−3	26	100	0	0
CN−	Not detected
VPs	Not detected
S2−	7 × 10−3	0.529	0.066	26	100	1	3.85
SO42−	1	437	34.56	26	100	1	3.85
Cl−	11	3010	207	26	100	3	11.54
NO3−	0.83	3.17	2.1	26	100	0	0
TFe	0.52	16	3.72	26	100	22	84.62
Mn	8×10−3	0.956	0.132	26	100	1	3.85

, MnO₄²⁻, NH₃-N, TP, Cu, Zn, F⁻, As, Pd, S²⁻, SO₄²⁻, Cl⁻, NO₃⁻, TFe, and Mn, were detected in all 26 sample groups collected during the wet season, resulting in RNSFs of 100%. Two factors, Hg and Cd, were detected only in 3 and 1 sample groups, respectively, yielding RNSFs of 11.54% and 3.85%. Among the 16 identified factors, eight factors exhibit measured concentrations that exceed the classification limits of grade III (see Table 4). They are MnO₄²⁻, NH₃-N, TP, S²⁻, SO₄²⁻, Cl⁻, TFe, Mn. NOMs for these factors are 24, 16, 15, 1, 1, 3, 22, 1, and the corresponding RNOMs are 92.31%, 61.54%, 57.69%, 3.85%, 3.85%, 11.54%, 84.62% and 3.85%, respectively.

4.2. Data Reliability Results

By substituting molar concentrations of the main anions and cations in water samples into Equation (2), relative errors of the charge balance for water samples collected during the dry season were calculated to be between −4.91% and −4.09% (see Figure 4). In contrast, the relative errors of charge balance for water samples in the wet season ranged between −3.59% and −4.24% (see Figure 4), both of which satisfy the requirement of being less than 5% in absolute value.

By substituting milligram equivalents of Ca²⁺ and Mg²⁺ and total hardness in water samples into Equation (4), the relative errors of the total hardness balance for water samples collected during the dry season were calculated to be between −0.90% and −0.87% (see Figure 5), while those for water samples in the wet season were between −0.44% and −0.87% (see Figure 5). Both of these values are below the 1% limit in absolute value.

4.3. Evaluation Results

4.3.1. Temporal Variations of SWQ

Generally, SWQ in Zhuxihe River is poor (see

Table 5. SWQ classifications during Dry and Wet Seasons at Sampling Sites in the Zhuxihe River Basin.

Sampling Sites	Dry Season		Wet Season
	Grade	FRLs	Grade	FRLs
SW1	III	NH3-N, TP	IV	MnO42−
SW2	IV	NH3-N	IV	MnO42−
SW3	III	NH3-N, TP	IV	MnO42−
SW4	III	Hg	III	MnO42−
SW5	V	TP	V	MnO42−
SW6	V	TP	IV	MnO42−
SW7	V	TP	IV	MnO42−, TP
SW8	IV	MnO42−, NH3-N	V	TP
SW9	V	TP	V	TP
SW10	V	TP	IV	MnO42−, TP
SW11	IV	MnO42−, TP	IV	MnO42−, TP
SW12	III	MnO42−, TP, Hg	V	TP
SW13	V	MnO42−	V	TP
SW14	IV	MnO42−, TP	V	TP
SW15	III	MnO42−, TP, Hg	V	TP
SW16	III	MnO42−, Hg	IV	MnO42−
SW17	III	MnO42−, Hg	IV	MnO42−, TP
SW18	IV	MnO42−	IV	MnO42−, NH3-N, TP
SW19	III	MnO42−, TP, Hg	IV	MnO42−
SW20	III	Hg	IV	MnO42−, TP
SW21	III	TP, Hg	IV	MnO42−, NH3-N
SW22	III	Hg	IV	MnO42−, TP
SW23	V	TP	IV	MnO42−, TP
SW24	V	TP	IV	MnO42−
SW25	V	MnO42−, NH3-N, TP	V	TP
SW26	IV	MnO42−	II	MnO42−, TP

). FRLs in Table 5 represent factors reaching limits that contribute to the SWQ classifications. As shown, the factors MnO₄²⁻, NH₃-N, and TP significantly contribute to the deterioration of water quality grades.

(1)

Dry season results

In the dry season, MnO₄²⁻ influences 19.23% of sites classified as grade III, 19.23% as grade IV, and 7.69% as grade V, affecting a total of 46.15% of sites within grades III to V. NH₃-N impacts 7.69% of sites classified as grade III, 7.69% as grade IV, and 3.85% as grade V, resulting in a total effect on 19.23% of sites in grades III to V. TP affects 23.08% of sites classified as grade III, 7.69% as grade IV, and 30.77% as grade V, with an overall impact on 61.54% of sites in grades III to V.

(2)

Wet season results

In the wet season, MnO₄²⁻ affects 3.85% of sites classified as grade III, 61.54% as grade IV, and 3.85% as grade V, totalling 69.24% of sites within grades III to V. NH₃-N has an effect on 7.69% of grade IV sites, with no impact on grades III and V. TP influences 30.78% of sites classified as grade IV and 26.92% as grade V, but has no effect on grade III sites, affecting a total of 50.7% of sites in grades IV to V.

4.3.2. Spatial Distribution of SWQ

(1)

Distribution of SWQ During Dry Season

Figure 6 presents the results of the synthetic evaluation of SWQ based on 26 groups of water samples collected during the dry season. The sampling sites are marked using the colors green, yellow, and red, which correspond to water quality falling into grades III, IV, and V, respectively. As shown, SWQ of the Zhuxihe River during the dry season was predominantly classified within grades III, IV, and V, with no samples falling into grades I or II. Specifically, grade III comprises 11 groups, grade IV includes 6 groups and grade V accounts for 9 groups.

SWQ, classified as grade III, accounts for 42.31% of the total samples and is distributed in the downstream section of the main river in Puqian Town, the upstream section of the main river in Fengpo Town, and along the Jiaolonggou Ditch. SWQ, classified as grade IV, constitutes 23.08% of the total samples. These sites are distributed in the midstream section of the main river, which flows through Jinshan Town, in the northern part of Baimangxi Stream, and at the confluence of the main river and Shanggouxi Stream. The quality of surface water samples classified as grade V comprises 34.61% of the total samples. These sites are distributed in the upstream section of the main river in Fengpo Town and the downstream section of the main river in Puqian Town, and they are also scattered throughout the downstream section of the Baimangxi Stream.

(2)

Distribution of SWQ during the wet season

SWQ during the wet season is primarily classified into grades II, III, IV, and V, which are represented by the colors blue, green, yellow, and red, respectively (see Figure 7). As illustrated, the classifications from grades II to V consist of 1 group, 1 group, 16 groups, and 8 groups.

SWQ, classified as grades II and III, comprises only 7.7% of the total samples, with these sites located along the Jiaolonggou Ditch and Baimangxi Stream. SWQ, classified as grade IV, accounts for 61.53% of the total samples. These sites are distributed in the downstream section of the main river in Puqian Town, the upstream section of the main river in Fengpo Town, at the confluence of the main river and Paigangxii Stream, and at the confluence of the main river and Baimangxi Stream. Additionally, the quality of surface water samples classified as grade V constitutes 30.77% of the total samples. These sites are distributed in the midstream section of the main river in Jinshan Town, as well as scattered in the Baimangxi Stream, Shangzhouxi Stream, and the upstream section of the main river in Fengpo Town.

4.3.3. Temporal Variations of DWQ

DWQ in the Zhuxihe River is generally worse than the SWQ. As indicated in

Table 6. DWQ classifications during Dry and Wet Seasons at Sampling Sites in the Zhuxihe River Basin.

Sampling Sites	Dry Season		Wet Season
	Grade	FRLs	Grade	FRLs
SW1	IV	TFe	IV	MnO42−
SW2	IV	NH3-N	IV	MnO42−
SW3	IV	TFe	IV	MnO42−
SW4	III	Hg, TFe	IV	TFe
SW5	V	TP	V	MnO42−
SW6	V	TP	IV	MnO42−, TFe
SW7	V	TP	IV	MnO42−, TP, TFe
SW8	V	TFe	V	TP
SW9	V	TP	V	TP
SW10	V	TP	IV	MnO42−, TP, TFe
SW11	IV	MnO42−, TP, TFe	IV	MnO42−, TP, TFe
SW12	IV	TFe	V	TP
SW13	V	MnO42−	V	TP
SW14	IV	MnO42−, TP, TFe	V	TP
SW15	IV	TFe	V	TP
SW16	IV	TFe	V	TFe
SW17	V	Cl−	V	TFe
SW18	IV	MnO42−, TFe	V	TFe
SW19	IV	TFe	IV	MnO42−, TFe
SW20	V	Cl−	IV	MnO42−, TP, TFe
SW21	IV	Cl−	V	TFe
SW22	V	SO42−, Cl−	V	Cl−
SW23	V	TP, SO42−, Cl−	V	Cl−
SW24	V	TP, SO42−, Cl−	V	SO42−, Cl−
SW25	V	MnO42−, NH3-N, TP, TFe	V	TP
SW26	IV	MnO42−, TFe	IV	TFe

, MnO₄²⁻, TP, and TFe are the main contributors to the deterioration of DWQ.

(1)

Dry season results

During the dry season, TFe is the predominant contaminant, impacting half of the sampling sites classified as grades III, IV, and V. Specifically, elevated TFe levels are measured in 3.85% of grade III sites, 38.46% of grade IV sites, and 7.69% of grade V sites, totally accounting for 50% of the sites within these grades. MnO₄²⁻ totally affects 23.07% of sites classified as grades IV and V, with 15.38% as grade IV and 7.69% as grade V, respectively. TP impacts 30.77% of sites. Among them, 7.69% of sites are classified as grade IV and 23.08% as grade V.

(2)

Wet season results

During the wet season, TFe is also the most significant factor deteriorating water quality, affecting 46.15% of sites, with 30.77% classified as grade IV and 15.38% as grade V. MnO₄²⁻ impacts 38.46% of sites, where 34.62% are classified as grade IV and 3.85% as grade V. TP affects 38.46% of sites, with 11.54% classified as grade IV and 26.9% as grade V.

4.3.4. Spatial Distribution of DWQ

(1)

Distribution of DWQ during the dry season

The evaluation results indicate that DWQ in the dry season is primarily classified as grades IV to V, with only one sampling site located in the Jiaolonggou Ditch, classified as grade III (see Figure 8). Water quality classified as grade IV accounts for 46.15% of the total samples and is located in the upstream section of the main river in Fengpo Town, the midstream section of the main river in Jinshan Town, as well as scattered locations in Baimangxi Stream and Paigangxi Stream. Additionally, water quality classified as grade V constitutes 50% of the total samples, which are distributed in the upstream section of the main river in Fengpo Town, the downstream section of the main river in Puqian Town, and also scattered in Baimangxi Stream.

(2)

Distribution of DWQ during the wet season

As shown in Figure 9, No DWQ samples fall within grades I to III; all samples are classified as grades IV and V. Specifically, 11 groups of water samples are classified as grade IV, constituting 42.3% of the total samples. These samples are primarily located in the upstream section of the main river in Fengpo Town, with additional occurrences in Jiaolonggou Ditch, Baimangxi Stream, and the downstream section of the main river in Puqian Town. Water quality classified as grade V accounts for 57.7% of the total samples. This classification is mainly distributed in the midstream section of the main river in Jinshan Town and the downstream section of the main river in Puqian Town. It is also scattered throughout the upstream section of the main river in Fengpo Town, as well as in the Baimangxi Stream and Paigangxi Stream.

5. Discussion

5.1. Contribution of Factors to Water Quality Deterioration

The above results indicate that SWQ and DWQ vary between dry and wet seasons due to different influencing factors. For surface water, the main factors contributing to the quality deterioration include MnO₄²⁻, NH₃-N, and TP. For centralized domestic drinking water, the main factors affecting its quality are MnO₄²⁻, TP, and TFe (see Table 5 and Table 6).

For SWQ, MnO₄²⁻ is the most significant factor, affecting a total of 115.39% of sites across both seasons (see Figure 10). As shown, its impact is notably greater during the wet season (69.24%) compared to the dry season (46.15%), indicating that runoff and increased precipitation exacerbate its presence. TP is the second most influential factor, impacting 61.56% of sites in the dry season and 50.70% in the wet season, resulting in a cumulative effect of 112.26%. NH₃-N exhibits a moderate influence, with a total impact of 26.92%, and its effect is stronger during the dry season.

For DWQ, TFe emerges as the predominant factor, affecting nearly half of the sites in both the dry (50.00%) and wet (46.15%) seasons, leading to a cumulative effect of 96.15% (see Figure 10). TP also plays a critical role, impacting 33.77% of sites in the dry season and 38.46% in the wet season, contributing to an overall effect of 72.23%. The influence of MnO₄²⁻ is substantial as well, with a total impact of 61.53%, demonstrating a stronger effect during the wet season (38.46%) compared to the dry season (23.07%), mirroring its trend in SWQ.

5.2. Temporal and Spatial Characteristics of Main Contributing Factors

As shown in Figure 2 and Figure 3, measured concentrations of factors show different temporal and spatial variation characteristics. For analysis and discussion, we selected those that mainly contribute to the deterioration of water quality in the main river (see Figure 11).

(1)

Temporal variation

As shown in Figure 11, seasonal variations have a significant impact on the water quality of the Zhuxihe River, with most pollutants exhibiting higher concentrations during the wet season. MnO₄²⁻, a factor to reflect the pollution of organic and inorganic oxidizable substances in water, generally increases in the wet season, likely due to enhanced surface runoff that transports organic and inorganic matter into the river [33,34]. Similarly, concentrations of TP and NH₃-N tend to rise during this period, possibly indicating increased nutrient input from agricultural and domestic sources [35]. Additionally, TFe displays seasonal fluctuations, with certain areas experiencing elevated concentrations in the wet season, likely associated with the seasonal change of hydrodynamics conditions and sedimentary environment [36].

(2)

Spatial variation

Spatially, water quality varies along the main river, with different concentration levels measured in the upstream, midstream, and downstream sections (see Figure 11). As shown, the upstream section (Fengpo Town) and the midstream section (Jinshan Town) exhibit relatively low concentrations of TP and TFe, suggesting limited external influences. In contrast, the downstream section (Puqian Town) displays the most severe pollution, characterized by elevated levels of TP and TFe, which indicate cumulative pollution effects from upstream sources, as well as additional local inputs. The distributions of MnO₄²⁻ and NH₃-N concentrations along the main river show minimal fluctuations, predominantly aligning with grade III classification. This suggests that both organic and inorganic pollutants are prevalent throughout the main river [37].

5.3. Suggestions for Future Work

Based on the findings of this study, three suggestions are proposed to enhance the understanding of water quality issues and improve management strategies in the Zhuxihe River Basin.

(1)

Isotopic tracing for pollution source identification

Future research should consider employing isotopic tracing techniques to more accurately quantify and trace the sources of pollutants in water systems, which would facilitate a more precise identification of pollutant origins, especially those stemming from agricultural runoff, livestock farming, and aquaculture activities.

(2)

Long-term monitoring and data integration

A long-term monitoring program should be established that integrates both surface water and groundwater quality assessments, which would aid in tracking seasonal and annual trends, thereby providing more comprehensive data for evaluating the effectiveness of remediation efforts and offering a clearer picture of long-term water quality dynamics.

(3)

Integrated strategies for water quality improvement

Local government should prioritize the enhancement of integrated strategies aimed at improving water quality, which include exploring alternative farming methods, optimizing irrigation systems, and implementing nutrient management strategies to minimize fertilizer and pesticide runoff. Furthermore, it is crucial to evaluate the effectiveness of existing pollution mitigation measures, such as wastewater treatment and sediment control, to enhance long-term water quality. Additionally, ecological restoration initiatives, including wetland creation and the rehabilitation of riparian zones, should be investigated to improve pollutant retention and promote biodiversity. Finally, fostering stakeholder engagement is essential for effective water management, with a focus on involving local communities, farmers, and industries in collaborative and participatory approaches to water pollution control.

6. Conclusions

This study provides insights into the current state of water quality in the Zhuxihe River, examining the temporal and spatial characteristics of factors contributing to water deterioration.

Based on the findings, the following conclusions can be drawn:

(1)

SWQ is primarily affected by chemical factors, such as MnO₄²⁻, NH₃-N, and TP, with high levels of concentrations in both dry and wet seasons. Specifically, MnO₄²⁻ has the most significant impact during the wet season. Additionally, TP and NH₃-N exhibit seasonal fluctuations. This finding suggests that agricultural runoff and wastewater discharge might be reasons for water deterioration.

(2)

Evaluation results reveal that the water in Zhuxihe River is unsuitable for drinking purposes. DWQ is mainly affected by TFe, MnO₄²⁻, and TP, with TFe identified as the predominant pollutant. This finding suggests that the river’s hydrodynamic conditions and the accumulation of pollutants in sediments might contribute to the deterioration of water quality.

(3)

Spatial distribution analysis indicates that the downstream section of the river experiences higher levels of pollution, reflecting an increase in local inputs and the cumulative effects of upstream pollution sources.

(4)

Potential solutions, including phytoremediation, microbial remediation, and source reduction of pollutants, could be considered key strategies for addressing water quality issues in the Zhuxihe River. Furthermore, future research should focus on optimizing these techniques, assessing their long-term effectiveness, and exploring the integration of these strategies with other environmental management practices to achieve sustainable improvements in water quality.