Assessment of Water Quality Using Chemometrics and Multivariate Statistics: A Case Study in Chaobai River Replenished by Reclaimed Water, North China

Abstract

Dry rivers could be effectively recovered by reclaimed water in North China, while river water quality would be an important issue. Therefore, it is important to understand the spatiotemporal variation and controlling factors of river water. Water samples were collected during March, May, July, September, and November in the year 2010, then 20 parameters were analyzed. The water environment was oxidizing and alkaline, which was beneficial for nitrification. Nitrate was the main nitrogen form. Depleted and enriched isotopes were found in reclaimed water and river water, respectively. Total nitrogen (TN) and total phosphorus (TP) of reclaimed water exceed the threshold of reclaimed water reuse standard and Class V in the surface water quality criteria. Most river water was at the severe eutrophication level. The sodium adsorption ratio indicated a medium harmful level for irrigation purpose. Significant spatial and temporal variation was explored by cluster analysis. Five months and nine stations were both classified into two distinct clusters. It was found that 6 parameters (chloride: Cl⁻, sulphate: SO₄²⁻, potassium: K⁺, sodium: Na⁺, magnesium: Mg²⁺, and total dissolved solids: TDS) had significant upward temporal variation, and 12 parameters (dissolved oxygen: DO, electric conductivity: EC, bicarbonate: HCO₃⁻, K⁺, Na⁺, Ca²⁺, TDS, nitrite-nitrogen: NO₂-N, nitrate nitrogen: NO₃-N, TN, TP, and chlorophyll a: Chl.a) and 4 parameters (Mg²⁺, ammonia nitrogen: NH₃-N, and the oxygen-18 and hydron-2 stable isotope: δ¹⁸O and δ²H) had a significant downward and upward spatial trend, respectively. The Gibbs plot showed that river water chemistry was mainly controlled by a water–rock interaction. The ionic relationship and principal component analysis showed that river water had undergone the dissolution of carbonate, calcite, and silicate minerals, cation exchange, a process of nitrification, photosynthesis of phytoplankton, and stable isotope enrichment. In addition, gypsum and salt rock have a potential dissolution process.

1. Introduction

North China has been facing serious water resources shortage in recent decades, as a result of continual drought, large consumption of water resources, water pollution, and economic development [1,2]. The groundwater table continued to decline and many rivers have been cut off or dried up for years [3]. Beijing as the capital of China, i.e., a big city located in North China, has also been facing massive water shortage. Multi-year average precipitation and evaporation are about 590 mm and 1800 mm, respectively [4,5]. Surface water resources of Beijing was 7.22 × 10⁸ m³, and 14.32 × 10⁸ m³ in the year 2010 and 2018, respectively [4]. In 2018, the inflow to Beijing from the middle route of the South-to-North Water Diversion Project was 11.92 × 10⁸ m³. Consumption of Beijing water resources was 35.2 × 10⁸ m³ and 39.3 × 10⁸ m³ in 2010 and 2018, respectively. Then, the shortage gap was satisfied by excessive use of groundwater. At the same time, utilization of reclaimed water rapidly increased from 2010 (6.8 × 10⁸ m³) to 2018 (10.8 × 10⁸ m³) [4,5]. Now, reclaimed water has become the “stable second water source of Beijing”, which was used for industry reuse, agricultural irrigation, river and lake landscape, and municipal utility [6]. As reclaimed water originated from treated wastewater, water quality was significantly different from natural surface water. High and complex content of salinity, nutrients (nitrogen and phosphorus), metals, and organic matter were remarkable features of reclaimed water. This could lead to soil salination [7], accumulation of heavy metals [8], groundwater pollution risk [9,10], antibiotics risk [11], and nutrient load in surface water [12].

Therefore, an understanding of water quality is very important for better use of reclaimed water. At present, chemometrics and multivariate statistics could provide powerful exploration for revealing water chemistry/quality characteristics. The Gibbs plot was depicted by drawing the relationship of the major ion ratio vs. total dissolved solids (TDS) of water, which included major rivers, rainfall, and seawater in the world [13]. It was powerful to determine the controlling mechanisms, which contained natural processes (atmospheric precipitation, rock weathering, and evaporation–crystallization) and anthropogenic activities [14]. A complex interaction among lithosphere, atmosphere, hydrosphere, and biosphere usually caused lithological weathering, which was the source of river water chemistry [14,15,16]. Stoichiometric analysis, e.g., the relationship of different combinations of dissolved cations and anions, could provide qualitative sources of ions in river water, such as evaporites, carbonates, and silicates [17,18,19]. Multivariate statistical analyses are particularly useful to explore the water chemistry/quality data set. Correlation analysis is powerful to interpret the relationship of water quality data and to infer specific water chemical processes [20]. The inner characteristics and distribution rule of water quality could be explored using a cluster analysis, and the similarity and dissimilarity also can be clarified. It usually contains a cluster of water chemical variables and samples. Consequently, ionic transformation and spatiotemporal variation could be clearly delineated [21,22]. Analysis of variance can be used to identify whether the difference of water parameters is significant, furtherly the spatiotemporal variation could be ascertained quantificationally [23]. Water quality has an extent of random and uncertainty, which will increase the difficulty of understanding the data. By dimensionality reduction of a large amount of data, the principal component analysis could determine the key water quality parameters, which are used to identify pollutant sources and the transforming mechanism of water chemistry [24,25]. In the past, research of river water quality mainly focused on natural rivers or rivers polluted by different sources of pollutants. Less studies were performed on rivers mainly replenished by reclaimed water. Generally, the single research method was usually applied to explore the river water quality. In this study, we try to combine the application of chemometrics and multivariate statistics methods for river water quality. So, the problem of river water quality/chemistry, e.g., the relationship of different water quality parameters, spatiotemporal variation, and evolution of water chemistry, could be quantitatively solved. Therefore, the combination of chemometrics and multivariate statistics would be more effectively to clarify the water chemical composition and governing factors. This study would provide sufficient suggestions for reclaimed water reuse and management of river water quality, even for water quality control issues of discharge from wastewater treatment plants.

For recovery of the dry Chaobai River, reclaimed water with a flow of 1.0 × 10⁵ m³/d, treated from Wenyu River water using membrane bioreactor technology, was moved to Chaobai River for ecological implementation. So, we researched the water quality of reclaimed water and river water for better use of reclaimed water. The main objectives of our study are (1) to understand the physical and chemical composition of reclaimed water and river water; (2) to clarify the spatial and temporal variation of water parameters; and (3) to ascertain the governing factors of water chemical evolution in the Chaobai River replenished by reclaimed water.

2. Materials and Methods

2.1. Study Site

The study site, located on the northeast of Beijing city, is the water course in Shunyi County, belonging to Chaobai River (Figure 1). The climate here has a seasonal temperature with a semi-humid monsoon climate with four distinct seasons. The multi-annual average temperature, annual rainfall, and evaporation are 11.8 °C, 614.9 mm, and 1175 mm, respectively. Meanwhile, the precipitation was concentrated mostly from June to September [26].

Chaobai River has been dry as a result of continual dry weather and the impoundment of the Miyun Reservoir since 1999, which belongs to the Quaternary Holocene alluvial–diluvial strata. The riverbed is mainly composed of fine sand, silt, and gravel layers. The thickness of the sand layer is generally from ten to a few dozen meters. Floodplain and terraces on both banks of the river channel are loam and sand [26]. While, Wenyu River (Figure 1) near the Chaobai River always has water flow throughout the year, which is mainly composed of domestic wastewater from surrounding communities. Wastewater in Wenyu River was treated using membrane bioreactor technology (MBR). Consequently, treated wastewater (reclaimed water) was transported to the Jian River (length: 4 km and width: 50–90 m; tributary of Chaobai River) by a water pipeline since the year 2007 [27]. Then, the Chaobai River channel from dam B to dam D was replenished by reclaimed water with a flow of 2.5 m³/s. Meanwhile, river flows freely through dam C. Most reclaimed water was stored in this section of the river as a result of the dams, except for evaporation and infiltration. The main channel of the river has a length of 7.3 km, width of 200–400 m, and average water depth of 2.5 m. River water flow was also changed by the dams after the replenishment of reclaimed water. One direction was the original river flow, i.e., southeast (to SY09), the other was northeast (to SY04; Figure 1).

2.2. Methods

2.2.1. Water Sampling

For investigating the water quality and controlling factors of river water replenished by reclaimed water in the Chaobai River, samples of reclaimed water and river water along the river channel were collected during March, May, July, September, and November in 2010. Nine monitoring stations (SY01–SY09) are shown in Figure 1, which included one reclaimed water and eight river water stations. In addition, one sample of reclaimed water in March was missed. Sample bottles with different volumes (100 mL and 500 mL) were made of polyethylene. The bottles were cleaned once with detergent, then cleaned once with tap water, and finally cleaned with deionized water 3 times. Water samples were collected at a depth of 50 cm below the river water surface using a plexiglass water collector. Three bottles of the water sample were collected for each monitoring station every time. The polyethylene bottles were prerinsed with water samples three times, before the final water sample was collected. The sampling frequency was five times for one year, and the sampling time was from 9:00 am to 5:00 pm. Three bottles of samples were collected at each station. The 100 mL sample was used for the determination of water stable isotopes, and the cap of the bottle was sealed with tape to prevent evaporation. A sample of 500 mL for the determination of anions, cations, nitrogen, and phosphorus was used and another sample of 500 mL was used for chlorophyll a (Chl. a) determination. All samples were stored in a portable cooler containing ice packs under 4 °C.

The precipitation data was measured using the tipping bucket automatic rain sensor (CG-04-D1, Hebei Yiqing Electronic Technology Company, Handan, China) on the roof of the Geographical Science Museum of the Institute of Geographical Sciences and Natural Resources Research, Chinese Academy of Sciences (40°00′11″ N, 116°23′07″ E, 45 m above sea level, about 10 m above the ground) from January to December 2010. This sampling point was 28 km from the study area in a straight line. A total of 17 precipitation events were collected, with a total precipitation of 412.7 mm. Then, monthly distribution data of precipitation in 2010 was obtained by accumulating precipitation events into the month.

2.2.2. Analytical Techniques

pH, water temperature (T, °C), dissolved oxygen (DO), and electric conductivity (EC) were measured by the portable multi-parameter water quality analyzer (American Hach HQ-40d), which was produced by HACH Company (Loveland, CO, USA). Water samples were taken back to the laboratory under 4 °C cold storage and analyzed within 24 h. The bicarbonate (HCO₃⁻) was determined by titration under the addition of sulfuric acid (0.02 mol/L), which the endpoint of titration had methyl orange as an indicator. Before further ionic analysis, water samples were filtered through a 0.45 μm Millipore membrane. Major cations including potassium (K⁺), sodium (Na⁺), calcium (Ca²⁺), and magnesium (Mg²⁺) were measured by inductively coupled plasma spectroscope (ICP-OES Optima 5300DV), produced by Perkinelmer Instruments Co., LTD (Norwalk, Connecticut, USA), with a detection limit of 0.01 mg/L. Major anions including chloride (Cl⁻), sulphate (SO₄²⁻), and nitrate (NO₃⁻) were measured by ion chromatograph (Thermo Fisher ICS2100) produced by the DIONEX company (Sunnyvale, CA, USA), and the detection limit was 0.01 mg/L. Ammonia nitrogen (NH₃-N), nitrite nitrogen (NO₂-N), total nitrogen (TN), and total phosphorus (TP) were measured using AMS’s Smartchem 200 batch analyzer produced by Alliance company (Paris, France), with a detection limit of 0.01 mg/L. Stable isotopes (δ²H and δ¹⁸O) were measured by a laser spectroscopic instrument (LGR DLT-100, Los Gatos Research, Mountain View, CA, USA), with the standard of Vienna standard mean ocean water (VSMOW). The precisions of δ¹⁸O and δ²H were 0.2‰ and 0.6‰, respectively. TDS was measured by the gravimetric method. The filtered water sample (200 mL) was placed in an evaporating dish weighed to a constant weight, and then baked to a constant weight at 103 ~ 105 °C. The TDS value was calculated by the increased weight [28]. The collected chlorophyll a (Chl. a) sample samples were stored at 4 °C, and 1 mL of 1% magnesium carbonate suspension was added to each liter of water samples to prevent pigmentation caused by acidification. In the laboratory, samples were filtered and concentrated, and the filter membrane was fully grinded and extracted, then dissolved in acetone to a constant volume, and finally the supernatant was measured by spectrophotometry. These experimental details were referred from the analytical book [28].

2.3. Data Processing

Data of the major ions need to be balanced with an error of less than 5% before further analysis. The summary statistics (e.g., mean, max, min, and coefficient of variation) of water chemistry were performed by the descriptive statistics package in SPSS 16.0 software (International Business Machines Corporation, Armonk, NY, USA) [29]. Analysis of variance (ANOVA), hierarchical cluster analysis (HCA), and principal component analysis (PCA) were all performed by SPSS [29]. Analysis of variance needs two assumptions about the data, which includes normal distribution and homogeneity of variance, respectively. These two parameters could be judged by the coefficients of skewness and Kurtosis, and Levene. The data could be considered to obey the hypothesis of normal distribution and homogeneity of variance, if the p value was more than 0.05 (p > 0.05). PCA is a way of selecting factors belonging to the factor analysis. Whether the data are suitable for factor analysis, it could be identified by KMO (Kaiser–Meyer–Olkin) and its significance (sig.) level test. If the KMO test value is greater than 100 and p < 0.05, then PCA could be performed. Respect to hierarchical clustering, especially for R-type, the data need to be standardized for avoiding the difference of dimension and orders of magnitude for variables [29]. The normality of water quality data was checked firstly by SPSS 16.0 software [29]. Then, all the data (except pH) were log-transformed and standardized before further analysis [30,31].

In addition, index, i.e., $\overline{R^2}$ was used for judging the selection of key variable in HCA, that was calculated as follows:

$$\overline{R^2}=\frac{{\displaystyle \sum {\mathrm{r}}^2}}{m-1}$$

(1)

where, ${\mathrm{r}}^2$ represents the correlation coefficient between different variables in clusters, and m represents the number of variables in one cluster.

Meanwhile, PHREEQC (Version 3, United States Geological Survey) is a hydrogeochemical simulation software based on C language [32]. It is mainly used to solve the analysis of chemical components, solute transport, and dynamic chemical reactions in the interaction system of water, gas, and rock–soil. The saturation index (SI) of major minerals was calculated by PHREEQC software.

Excessive sodium and salinity in irrigation water would result in sodium hazard. Calcium and magnesium in soil could be replaced by sodium, which leads to the reduction of permeability and soil harden [33]. Sodium adsorption ratio (SAR) calculated based on chemical variables was used to assess irrigation water quality, which was an effective evaluation index [34].

$$SAR=N{\mathrm{a}}^{+}/\sqrt{(C{a}^{2+}+M{g}^{2+})/2}$$

(2)

where, ionic concentrations are expressed in milliequivalent per liter (meq/L).

3. Results and Discussion

3.1. Water Chemical Composition

Physical and chemical compositions of water samples in the Chaobai River are given in

Table 1. Water chemical composition of reclaimed water and river water in the Chaobai River (March–November 2010).

	Reclaimed Water					River Water
	Min.	Max.	Mean	SD	C.V. (%)	Min.	Max.	Mean	SD	C.V. (%)
pH	7.65	8.21	7.93	0.25	3.16	8.08	9.45	8.81	0.41	4.68
T (°C)	6.90	28.60	21.03	9.65	45.92	1.90	29.70	17.42	10.30	59.12
DO (mg/L)	4.73	8.50	7.33	1.75	23.84	1.25	10.30	6.20	2.50	40.24
EC (μS/cm)	937	1025	982	36.88	3.75	935	1047	859	90.72	10.56
Cl−(mg/L)	72.90	114.00	87.65	18.65	21.27	65.2	148.0	98.15	19.96	20.34
HCO3−(mg/L)	219.00	299.00	261.75	33.24	12.70	230	386	227.78	49.01	21.52
SO42−(mg/L)	81.90	99.50	89.33	7.97	8.92	64.30	120.00	87.04	13.66	15.69
K+(mg/L)	12.90	18.60	16.25	2.83	17.44	6.32	25.80	15.25	4.82	31.60
Na+(mg/L)	80.20	108.00	94.45	11.62	12.30	57.3	135.0	88.46	17.51	19.79
Ca2+(mg/L)	60.10	66.40	63.47	3.17	5.00	23.7	75.2	48.92	12.80	26.17
Mg2+(mg/L)	22.60	28.60	26.00	2.71	10.42	23.3	39.9	29.10	3.43	11.80
NH3-N(mg/L)	0.03	0.11	0.08	0.03	43.92	0.02	1.81	0.38	0.42	109.57
NO2-N(mg/L)	0.22	0.76	0.38	0.26	67.66	0.00	0.74	0.16	0.15	91.60
NO3-N(mg/L)	10.60	19.40	14.75	3.83	25.99	0.03	19.90	7.03	5.95	84.66
TN (mg/L)	11.50	22.10	17.18	4.47	26.01	2.30	20.00	9.62	5.29	54.94
TP (mg/L)	0.44	1.52	1.05	0.46	43.8	0.09	2.76	0.60	0.55	92.00
Chl.a (μg/L)	0.50	3.58	2.03	1.77	87.03	1.96	175.00	58.87	45.18	76.74
TDS (mg/L)	568	655	601	39.35	6.55	419	794	540	80.07	14.82
δ18O (‰)	−8.55	−7.25	−7.99	0.63	−7.91	−8.64	−5.14	−6.85	0.98	−14.34
δ2H (‰)	−61.34	−55.36	−58.39	2.73	−4.68	−63.64	−43.70	−53.48	5.28	−9.86

. pH ranged from 7.65 to 9.45, with an average value of 8.37, which showed reclaimed water and river water were all alkaline. The average value in river water was higher than that in reclaimed water. Water temperature was mainly controlled by the operation of the wastewater treatment plant (WWTP) and air temperature of seasonal variation. Electric conductivity (EC) and TDS were the similar parameters indicating the total dissolved solids in aqueous solution, which in reclaimed water were both slightly higher than in river water. The order of nitrogen forms in reclaimed water and river water was: NO₃-N > NO₂-N > NH₃-N, and NO₃-N > NH₃-N > NO₂-N, respectively. The coefficient of variation (C.V.) of NH₃-N and NO₂-N was much higher than others. It indicated that they were prone to chemical transformation in the water environment. Considering the average, higher percentage of NO₃-N /TN in reclaimed water (85%) and river water (78%) indicated nitrate was the major nitrogen form. The order of anions in reclaimed water was: HCO₃⁻ > SO₄²⁻ > Cl⁻; while, the order in river water was: HCO₃⁻ > Cl⁻ > SO₄²⁻. Whereas, the order of cations (Na⁺ >Ca²⁺ >Mg²⁺ > K⁺) was the same in both waters. In terms of the average, seven indicators of river water were higher than in reclaimed water, which contained pH, Cl⁻, Mg²⁺, NH₃-N, Chl.a, δ¹⁸O, and δ²H. While, the remaining indexes were the opposite. A reduction of nutrients (NO₃-N, NO₂-N, TN, and TP) may be caused by the consumption of phytoplankton or dilution [35,36]. External input from surface runoff in the rainy season could lead to the increased content of Cl⁻ and Mg²⁺ [37,38]. Increased ammonia content may be caused by the mineralization of organic nitrogen [39,40]. High value of Chl.a indicated the reproduction of phytoplankton, e.g., alga in river water [36]. Stable isotopes (δ¹⁸O, δ²H) were depleted in reclaimed water, while enriched in river water (Table 1), and the similar enrichment phenomenon was also found in Huai River [10].

A Pearson correlation analysis was applied to explore the relationship of water chemical parameters by SPSS, and the results are given in

Table 2. Correlation of water chemical parameters in water samples in the Chaobai River (March–November 2010).

	pH	T	DO	EC	Cl−	HCO3−	SO42−	K+	Na+	Ca2+	Mg2+	TDS	NH3-N	NO2-N	NO3-N	TN	TP	Chl.a	δ18O	δ2H
pH	1.00
T	0.10	1.00
DO	0.16	0.37 *	1.00
EC	−0.18	0.09	0.16	1.00
Cl−	0.37 *	0.00	0.11	0.10	1.00
HCO3−	−0.25	0.14	0.40 **	0.66 **	0.21	1.00
SO42−	0.29	−0.15	−0.10	0.24	0.71 **	−0.04	1.00
K+	0.15	−0.08	0.38 *	0.59 **	0.59 **	0.66 **	0.51 **	1.00
Na+	0.11	0.06	0.27	0.70 **	0.60 **	0.59 **	0.61 **	0.93 **	1.00
Ca2+	−0.37 *	−0.04	0.42 **	0.77 **	−0.20	0.82 **	−0.16	0.51 **	0.46 **	1.00
Mg2+	0.08	0.03	0.21	−0.21	0.58 **	0.27	0.01	0.21	0.11	0.11	1.00
TDS	−0.05	−0.06	0.31 *	0.77 **	0.49 **	0.73 **	0.43 **	0.89 **	0.89 **	0.72 **	0.20	1.00
NH3-N	−0.26	0.07	−0.13	−0.35	−0.20	−0.06	−0.20	−0.23	−0.24	−0.15	0.07	−0.21	1.00
NO2-N	−0.07	0.19	−0.02	0.45 **	−0.17	0.34 *	−0.04	0.22	0.28	0.25	−0.36 *	0.23	−0.07	1.00
NO3-N	−0.15	−0.20	0.24	0.85 **	0.10	0.50 **	0.20	0.64 **	0.64 **	0.75 **	−0.20 **	0.77 **	−0.41 **	0.28	1.00
TN	−0.02	0.08	0.30	0.83 **	0.01	0.50 **	0.15	0.54 **	0.57 **	0.64 **	−0.37	0.63 **	−0.48 **	0.41 **	0.88 **	1.00
TP	−0.14	0.30	0.33	0.77 **	0.26	0.67 **	0.22	0.56 **	0.62 **	0.62 **	−0.02	0.70 **	−0.30	0.26	0.62 **	0.67 **	1.00
Chl.a	0.60 **	0.14	0.01	−0.44 **	0.25	−0.34 **	0.12	−0.03	−0.08	−0.53 **	0.15	−0.27	−0.09	−0.16	−0.38	−0.26	−0.29	1.00
δ18O	0.33	0.31	0.01	−0.70 **	0.29	−0.28	0.09	−0.19	−0.26	−0.67 **	0.36	−0.44 **	0.26	−0.21	−0.79 **	−0.63 **	−0.41 **	0.45 **	1.00
δ2H	0.32	0.23	−0.03	−0.77 **	0.31	−0.39	0.13	−0.27	−0.33	−0.71 **	0.34	−0.48 **	0.30	−0.28	−0.80 **	−0.66 **	−0.46 **	0.44 **	0.94 **	1.00

*: Correlation is significant at p < 0.05; **: Correlation is significant at p < 0.01. The observation includes samples of nine stations in five months except SY01 (four months), and the numbers are 44 (n = 44).

. Carbonate was observably positively correlated with K⁺, Na⁺, and Ca²⁺, which indicated the dissolution of minerals [41]. Chloride was prominently and positively related with SO₄²⁻, K⁺, Na⁺, Mg²⁺, and TDS. TDS was significantly and positively correlated with EC, Cl⁻, HCO₃⁻, SO₄²⁻, K⁺, Na⁺, Ca²⁺, NO₃-N, TN, and TP, respectively. Meanwhile, the same correlation was also found between EC with these ions. It indicated these ions were the major composition of TDS and EC. While, TDS were significantly and negatively correlated with stable isotopes and Chl.a. This may be due to the consumption of nitrogen and phosphorus by phytoplankton [35,36]. TN was significantly and positively related with TP, indicating their common origin. At the same time, they were both significantly and positively related with EC, HCO₃⁻, K⁺, Na⁺, Ca²⁺, TDS, and NO₃-N. In addition, TN was also positively and negatively significantly correlated with NH₃-N and NO₂-N, respectively. TN, TP, and other cations were mainly from reclaimed water, which could explain their significant relationship. The major nitrogen forms (NO₃-N) were furtherly confirmed by its significant and positive relationship with TN (Table 1 and Table 2). Part of NH₃-N will be released into the atmosphere [42,43,44], which could be inferred by its significant and negative relationship with TN.

Chl.a was remarkably and negatively related with EC, HCO₃⁻, Ca²⁺, δ¹⁸O, and δ²H. While, it was significantly and positively related with pH. CO₂ in river water could be converted into organic matter by the photosynthesis of alga, and HCO₃⁻ was the carbon form in the carbonate system [45,46]. While, the consumption of nitrogen, phosphorus, and carbon would lead to the decrease of EC. In addition, the amount of O₂ produced by the photosynthesis of alga was much greater than the one required for respiration, which increased DO content in the water [47,48]. Consequently, the reduction of HCO₃⁻ and the increase of DO would together raise the pH value (Table 1) [49,50]. Stable isotopes were both dramatically and negatively correlated with EC, Ca²⁺, TDS, NO₃-N, TN, and TP. Significant and positive correlation of δ¹⁸O and δ²H was determined by a stable isotope fractionation mechanism [51,52,53].

3.2. Spatial and Temporal Variation

3.2.1. pH, T, DO, EC, and TDS

Spatial and temporal variation of pH, T, DO, EC, and TDS is given in Figure 2. A first gradual upward and then downward trend were the clear spatial variation of pH, and the rising stage mainly occurred during stations from SY04 to SY06. The lowest pH was found in reclaimed water i.e., water sources. Moreover, pH in September was lower than in other months. Meanwhile, the pH value was very close in other months. River water received more replenishment of precipitation and surface runoff with low pH with range of 4.35–5.70 [54]. Obvious temporal variation of water temperature was found, that was mainly affected by air temperature. Its order was: July > May > September > March > November. Microbial activity was strongly influenced by water temperature. Consequently, the element cycle driven by a microbe, e.g., nitrogen would be affected by temperature variation [55]. Lower DO content was found in November, and a higher value was found in May and July (Figure 2c), while DO in other months was close. DO content varied drastically among monitoring stations. EC and TDS had the similar downward trend of spatial variation (Figure 2d,e). They both were higher in reclaimed water, and gradually decreased along the river. Furthermore, the lowest EC value (718 μS/cm) in March was found at one end of the river flow (SY04), and EC was close in different months. However, TDS in March, May, and July were slightly higher than that in September and November. Similar temporal variation of water temperature was also found in the Yongding River [56], while the spatial variation of these five parameters was different, due to the diverse water quality of reclaimed water [56] or treated wastewater [57], hydraulic conditions, and the geomorphic feature [58].

3.2.2. Major Cations and Anions

Figure 3 shows the spatial and temporal variation of major cations and anions. Cl⁻ has apparent temporal variation, and its order among months was: May > March > July > November > September (Figure 3a). Cl⁻ content increased gradually along the river water flow direction, except reclaimed water in March. Elevated chloride indicated the external input, e.g., dissolution of soil and/or sediment, or surface runoff, as chloride was the conservative ion [59]. The average value in this study was much less than the mean value in Kakoba sewage effluents (833.33 mg/L) [58]. While, the temporal variation was not significant [58]. The content of SO₄²⁻ was high in May, and low in July and September (Figure 3b). Meanwhile, it varied greatly among monitoring stations. Similar spatiotemporal variation of K⁺ and Na⁺ was found in Figure 3c,d. Their high values were found in March, while low values in September and November, and the medium in May and July. Their downward trend of spatial variation may be caused by ion exchange and/or dilution [19]. Variation of Ca²⁺ showed a remarkable downward spatial trend (Figure 3e). High Ca²⁺ content was firstly found in reclaimed water. However, its concentration gradually decreased along river flow, especially in SY04 and SY07 in May. The observed decrease may be the result of calcium precipitation, and/or calcium ions exchange with soils/minerals [17,60]. The content of Ca²⁺ was high in March and July, and low in May, while medium in other months. Gradual rising was the primary feature of spatial variation of Mg²⁺, except a high value of reclaimed water in March (Figure 3f). External input and/or dissolution of mineral with magnesium may contribute to an increase [61]. High and low value of Mg²⁺ was found in March and November, respectively. While, the medium one was in other months. Major cations and anions in the river impacted by sewage effluents in the Mediterranean had significant spatiotemporal variation [57]. Complexity and high variability occurred in different rivers with diverse conditions.

3.2.3. Nitrogen, Phosphorus, and Chl.a

Spatial and temporal variations of nitrogen, phosphorus, and Chl.a are given in Figure 4. Spatiotemporal variation of NH₃-N was obvious (Figure 4a), with the highest and the second highest value in SY01 (September) and SY04 (May), respectively. NH₃-N content in May and September was higher than in other months, while most values of NH₃-N were lower than 0.5 mg/L. Most NO₂-N content was lower than 0.4 mg/L (Figure 4b). Generally, spatial variation of NO₂-N was decreasing except for four peak values (SY01 and SY03 in November and SY04 and SY05 in July). These nitrogen forms also exhibited very high variability Rwizi River (Uganda) [58]. Gradual downward spatial variation of NO₃-N, TN, and TP was apparent. The evident features of NO₃-N and TN were the slight increase from SY05 to SY06 and sharp decrease in SY04. A gradual decrease was also found in the Sand River (Limpopo, South Africa) impacted by sewage effluents, and this may indicate the self-purification capacity of the river [62]. The peak value of TP was found in SY02 (May), which was higher than in reclaimed water. It may be due to the release of sediments and/or sudden external input [63,64]. While, TP in the Sand River fluctuated across the different sites [62]. First an increase and then a decrease were the feature of Chl.a spatial variation (Figure 4f). Meanwhile, it was higher downstream than upstream, with a peak value of SY08 in May (167 μg/L) and July (175 μg/L). High Chl.a was found in May and July, while a low value in November. There was a slightly low concentration of Chl.a in September corresponding to July, which may be caused by the dilution. Stations with a high peak value of Chl.a were the corresponding stations with low nitrogen and phosphorus. It further confirmed that the absorption and utilization of phytoplankton was the reason for nutrients reduction.

3.2.4. δ²H and δ¹⁸O

Figure 5 shows the spatiotemporal variation and the relationship of δ²H and δ¹⁸O in reclaimed water and river water. δ²H and δ¹⁸O had the similar variation (Figure 5a,b). The clear spatial trend was increasing variation of δ²H and δ¹⁸O from reclaimed water to the river channel end. It also indicated the enrichment process of stable isotopes, and the depleted and enriched isotopes in reclaimed water and the two ends (SY04 and SY09), respectively. Stable isotopes in March and May were higher, and gradually increased from July to November. In the dry season, e.g., March and May, high air temperature and less rainfall contributed to isotope enrichment fractionation. However, stable isotopes of river water would be prone to being depleted as a result of rainfall and surface runoff [65]. According to our observation, the precipitation in July was the largest (124.6 mm), followed by August (101.8 mm), which accounted for 30.19% and 24.67% of the annual precipitation respectively. During the rainy season (July–September), the cumulative precipitation was 297.9 mm, accounting for 72.18%. Therefore, isotopes in November also would be affected, as it was just after the rainy season. Almost all samples were located below the global meteoric water line (GMWL) and local meteoric water line (LMWL), except several samples on and near the LMWL line (Figure 5c), which indicated that most reclaimed water and river water were influenced by strong evaporation [52]. Additionally, it was consistent with the results of the isotope enrichment feature (March and May) in Figure 5a,b. Samples located on LMWL showed their source of atmospheric precipitation and less evaporation [53,66]. The evaporation line of stable isotopes among different months were different, and the order of the slope was: July (6.13) > March (5.72) > November (5.21) > May (4.85) > September (4.02).

3.3. Processes Controlling Water Chemistry

3.3.1. Gibbs Plot and Bivariate Plot

Mechanisms controlling river water chemistry included atmospheric precipitation, rock weathering, evaporation–crystallization processes [13], and anthropogenic activities [14]. The high ratio of Na⁺/(Na⁺+Ca²⁺), low ratio of Cl⁻/(Cl⁻+HCO₃⁻), and medium TDS concentration were found in the Gibbs plot (Figure 6). Meanwhile, distributions of reclaimed water and river water samples in cation and anion plots were both located in the rock dominance area. It indicated that river water chemistry was mainly controlled by reclaimed water or the interaction of river water with soil/rock. While, the effect of precipitation and evaporation was weak. The discharge of reclaimed water from WWTP was about 1.0 × 10⁵ m³/day. As a result, the TDS load was about 60.1 Ton/day, according to the average value of the TDS concentration (601 mg/L). Similar Gibbs plot results were found in the Yongding River, which was also replenished by reclaimed water [56]. Water chemistry in headwaters of the Yangtze River [15] and Yellow River [68] in China was governed by rock weathering, which was less impacted by human activities. While, evaporation–crystallization plays an important role in water chemistry in the river of an arid watershed, e.g., Northern Xinjiang, China [16].

Mineral dissolution, precipitation, and redox reaction in the water environment could be inferred by the relationship of different dissolved ions in waters [13,69]. The ratio of Ca²⁺/Mg²⁺ could indicate the dissolution of carbonate minerals [69]. The ratio value was about equal to 1, indicating the dissolution of dolomite, that included SY04 (March and July), four samples in November (SY05, SY06, SY07, and SY08), SY09 (July, September, and November; Figure 7a). The value was between 1 and 2, indicating the dissolution of calcite, which included reclaimed water (SY01), river water (SY02 and SY03) river samples of SY05 (March and May) and SY06 (Figure 7a). The ratio value of the remaining samples was less than 1, showing the decrease of Ca²⁺, which may be caused by ion exchange. Values of Ca²⁺+Mg²⁺ vs. the cation of all reclaimed water and river water samples were located above the equilibrium line (1:1; Figure 7b). It indicated that Ca²⁺ and Mg²⁺ were mainly from the dissolution of carbonate rock and calcite [18,41]. All samples of Na⁺+K⁺ vs. Cl⁻ (Figure 7c) were located below the equilibrium line, indicating that sodium and potassium ions were also affected by the dissolution of silicate minerals except the dissolution of salt rock [68]. Samples of Na⁺+Ca²⁺ vs. HCO₃⁻ (Figure 7d) were all located below the equilibrium line, showing that sodium and calcium were more than bicarbonate, which indicated the dissolution of calcium-bearing minerals. Most of SO₄²⁻+Cl⁻ vs. HCO₃⁻ were located below the equilibrium line (Figure 7e), indicating the influence of strong evaporation. Some of SO₄²⁻+HCO₃⁻ vs. Ca²⁺+Mg²⁺ were located near the line (Figure 7e), indicating the dissolution of carbonate minerals [70]. While, most were below the line (Figure 7e), indicating the dissolution of the silicate mineral. The same silicate weathering location was found between the plot of Ca²⁺+/Na⁺ vs. Mg²⁺/Na⁺ (Figure 7g), and plot of Ca²⁺/Na⁺ vs. HCO₃⁻/Na⁺ (Figure 7h), indicating the weathering and dissolution of the silicate mineral.

Samples of 1/2HCO₃⁻ + SO₄²⁻ vs. Ca²⁺ + Mg²⁺ were located near the equilibrium line (Figure 8a), indicating the dissolution of calcite, dolomite, and gypsum [71,72]. All samples were below the line except one sample, showing the excess Ca²⁺+Mg²⁺. These cations would be balanced by other anions, which may be silicate. Variation of the ratio of Na⁺/Ca²⁺+Mg²⁺ (Figure 8b) was first to increase and then to decrease, which indicated the occurrence of cation exchange with clay mineral or soils [70]. Lower values were found in SY08 and SY09 stations. The process of weathering and hydrolysis of carbonate rock or silicate minerals produced equal amounts of divalent cations, HCO₃⁻ and SO₄²⁻. As a result, Ca²⁺+Mg²⁺-HCO₃⁻-SO₄²⁻, and Na⁺-Cl⁻ were used to indicate the participation of cation exchange, respectively [71]. Some samples of Ca²⁺ + Mg²⁺- HCO₃⁻-SO₄²⁻ vs. Na⁺-Cl⁻ were located around the line (y = −x; Figure 8c), which indicated the clear cation exchange. Others showed the excess sodium ions from reclaimed water. In Figure 8d, the ratio of Na⁺/Cl⁻ samples was all more than 1, except three samples (SY08 and SY09 in November and SY09 in July). The average value of Na⁺/Cl⁻ in reclaimed water was 1.69, indicating the excess sodium. Values of Na⁺/Cl⁻ in most river water samples were close to or less than the reclaimed water, which indicated the external chloride ion. The ratio of Na⁺/Cl⁻ was greater than 1, indicating the possible cation exchange [73]. The downward trend of the Na⁺/Cl⁻ value may be caused by the following reason. A higher Na⁺ content exceeded the equilibrium concentration of exchangeable cations in the medium. Then, the exchange of Ca²⁺ and adsorbed Na⁺ would be suppressed, and even reverse cations exchange would occur. Na⁺ and Ca²⁺ in the water body exchange and Mg²⁺ may also participate in the exchange [17,19].

Saturation index (SI) of reclaimed water and river water samples is given in

Table 3. Saturation index (SI) of samples in monitoring stations in the Chaobai River (March–November 2010).

	SY01	SY02	SY03	SY04	SY05	SY06	SY07	SY08	SY09
CaSO4	−1.99	−1.97	−2.04	−2.23	−2.13	−2.12	−2.23	−2.20	−2.17
CaCO3	0.63	0.94	1.19	1.09	1.36	1.46	1.16	1.00	1.12
Ca Mg (CO3)2	1.16	1.76	2.26	2.32	2.73	2.89	2.47	2.12	2.32
CaSO4·H2O	−1.76	−1.73	−1.81	−2.00	−1.90	−1.89	−1.19	−1.96	−1.94
NaCl	−6.68	−6.58	−6.62	−4.39	−6.62	−4.26	−6.67	−6.71	−6.72

The numbers of measurements of all monitoring stations were five, except SY01 station (n = 4).

. Potential dissolution and precipitation processes in the aqueous solution could be inferred by SI values. A zero SI value indicated the equilibrium state of the mineral to the aqueous, and a positive value showed the supersaturated state, while a negative value indicated the unsaturated state [32]. The SI value of gypsum (CaSO₄ and CaSO₄·2H₂O) and halite (NaCl) was negative, while the SI value of calcite (CaCO₃) and dolomite (Ca Mg (CO₃)₂) was positive. This indicated the potential dissolution process of gypsum and salt rock, and precipitation of calcite and dolomite in reclaimed water and river water. The excessive Na⁺ of water samples may require silicate dissolution to balance it.

3.3.2. Redox Condition and Nitrogen Forms

Transformation and species of nitrogen in aqueous solution were strongly impacted by the redox environment [74,75], which could be characterized by DO and/or ORP measured in the field. The corresponding relationship of DO and three nitrogen forms (NO₃-N, NH₃-N and NO₂-N) is given in Figure 9. Nitrite was usually the intermediate of the nitrification reaction, which was not stable and easy to be oxidized [76]. In Figure 9, average DO content of river water in March, May, July, September, and November was 7.99 ± 2.07 mg/L, 5.64 ± 2.79 mg/L,7.90 ± 0.85 mg/L, 6.21 ± 1.66 mg/L, and 3.27 ± 1.41 mg/L, respectively. The saturated oxygen content of water was about 10 mg/L at normal temperature and atmospheric pressure [48,77]. DO content was more than 5 mg/L except the one in November. Additionally, the oxidizing environment contributes to nitrification [78]. The decrease in November probably was due to low temperature and less aquatic plants [49,79].

pH also played an important role in nitrogen transformation, and the double influences of pH and the redox environment could be evaluated by the pH–pE plot [78]. Therefore, samples of reclaimed water and river water were projected into Figure 10. All samples were located around the NO₃⁻ line, where the stable nitrogen form was nitrate (NO₃⁻, +5). High DO value of reclaimed water (7.33 ± 1.75) and river water (6.20 ± 2.47) showed the oxidizing water environment, which was beneficial for nitrate stable and nitrification [78,80].

3.3.3. Cluster Analysis and Spatiotemporal Similarity

For identifying the spatiotemporal variation, the key variables should be firstly confirmed. Hence, three cluster analysis results are shown in Figure 11. In Figure 11c, twenty water chemical parameters were divided into seven clusters by the R-type cluster method for variables. Clusters were as follows. I: δ¹⁸O, δ²H, pH, and Chl.a; II: Mg²⁺; III: Cl⁻, SO₄²⁻; IV: NH₃-N; V: T, DO; VI: K⁺, Na⁺, TDS, HCO₃⁻, Ca²⁺, NO₃-N, TN, EC, and TP; VII: NO₂-N. The R² value is calculated according to equation 1 if the quantity of parameters in one cluster was more than 2. The parameter with the highest R² value was retained for the key parameter. The R² value of δ¹⁸O and TDS was 0.45 and 4.80, respectively. As a result, nine parameters (δ¹⁸O, Mg²⁺, Cl⁻, SO₄²⁻, NH₃-N, T, DO, TDS, and NO₂-N) were selected as the key ones for further spatiotemporal cluster analysis. Five months were classified into two groups, Cluster I includes March and May, and Cluster II includes July, September, and November. These two distinct groups show the significant temporal variation. Spatial variation and similarity were analyzed by an HCA analysis (Figure 11c). All monitoring stations were classified into two groups, which include Cluster I (SY04, SY07, SY08, andSY09) and Cluster II (SY01, SY02, SY03, SY05, and SY06). Stations in Cluster II and Cluster I represent the upstream and downstream (Figure 1), respectively. The flow direction of reclaimed water along the river channel would eventually go in two directions, one was SY04 station and the other was SY09 station, because of the blocking effect of the rubber dam (Figure 1).

Parameters with a significant difference in different clusters were identified using an ANOVA analysis by SPSS software [29], and the corresponding results are given in

Table 4. Parameters in spatiotemporal clusters with a significant difference in the Chaobai River.

		Cl−	SO42−	K+	Na+	Mg2+	TDS
Temporal clusters	I	119.44 a	96.97 a	18.91 a	101.37 a	30.70 a	590.12 a
	II	83.19 b	81.12 b	13.10 b	81.21 b	27.63 b	518.00 b
		DO	EC	HCO3−	K+	Na+	Ca2+	Mg2+	TDS
Spatial Cluster	I	5.47 b	786 b	208 b	13.10 b	79.36 b	40.59 b	30.36 a	495 b
	II	7.00 a	940 a	250 a	17.21 a	97.04 a	58.22 a	27.51 b	588 a
		NH3-N	NO2-N	NO3-N	TN	TP	Chl.a	δ18O	δ2H
Spatial Cluster	I	0.51 a	0.124 b	2.04 b	5.33 b	0.281 b	71.00 b	−6.11 a	−49.20 a
	II	0.22 b	0.234 a	12.47 a	14.46 a	0.944 a	39.29 a	−7.66 b	−57.87 b

Note: Average values of the same parameters in corresponding groups with different letters are significantly different (p < 0.05). Temporal clusters/ Spatial Cluster

. Whereas, parameters with no significant difference were not listed. In temporal clusters, six parameters in Cluster I were significantly higher than the ones in Cluster II, which included Cl⁻, SO₄²⁻, K⁺, Na⁺, Mg²⁺, and TDS. In the spatial clusters, 12 parameters in Cluster I were significantly less than in Cluster II, which included DO, EC, HCO₃⁻, K⁺, Na⁺, Ca²⁺, TDS, NO₂-N, NO₃-N, TN, TP, and Chl.a. While, four parameters in Cluster I were significantly higher than in Cluster II, which included Mg²⁺, NH₃-N, δ¹⁸O, and δ²H. As a result, the parameters with a significant difference had remarkable spatiotemporal variation. δ¹⁸O and δ²H of reclaimed water (−7.99‰ ± 0.6‰ and −58.39‰ ± 2.73‰) were more depleted than the ones in downstream stations (SY04: −5.94‰ ± 0.58‰ SY07: −49.26‰ ± 4.68‰; SY09: −6.07‰ ± 0.66‰, and SY08: −49.15‰ ± 2.51‰). The evaporation process along the river may lead to more isotope enrichment. Reclaimed water was the main source of river water, besides the precipitation and surface runoff from precipitation. Stable isotopes of precipitation were more depleted compared with reclaimed water [10]. Therefore, strong evaporation was the controlling reason for stable isotope enrichment downstream [53,66].

3.3.4. Principal Component Analysis and Controlling Factors

For inferring the controlling factors, two clusters were further analyzed by PCA (Table S1), and the relationship plots are given in Figure 12. The principal components would be retained according to the corresponding eigenvalue >1 [24], and the critical related parameters could be kept in terms of an explaining proportion >0.60 [25]. In Cluster I, six principal components with a total explaining 90.46% were retained. PC1 had significant and positive loading with Cl⁻, SO₄²⁻, K⁺, Na⁺, TDS, δ¹⁸O, and δ²H. PC2 was highly and positively related with HCO₃⁻, Ca²⁺, and Mg²⁺. PC3 had a high and positive relation with EC, NO₂-N, and TP. High and positive loading with these three components showed main water chemical composition of TDS and EC, and mineral dissolution. PC4 had high and positive loading with δ¹⁸O and δ²H. PC5 was positively related with Na⁺, TN, and Chl.a, while negatively related with NH₃-N. PC6 was highly and positively related with NO₃-N and TN. NO₃-N was the main nitrogen formation formed by nitrification of NH₃-N [78]. As a result, the negative loading with NH₃-N could be found. Meanwhile, nitrate was also the nitrogen consumed by phytoplankton [35]. Therefore, PC5 and PC6 mainly indicated the nitrogen transformation and the photosynthesis by phytoplankton. In Cluster II, five principal components explaining 87% were retained. PC1 had high and positive loading with EC, Cl⁻, HCO₃⁻, K⁺, Na⁺, Ca²⁺, Mg²⁺, and TDS, which indicated reclaimed water was the main source of major ions. Additionally, these stations were strongly impacted by reclaimed water. PC2 was highly related with Chl.a, δ¹⁸O, and δ²H. PC4 had a negative loading with SO₄²⁻. There were no highly related parameters to PC3 and PC5. High and positive loading of δ¹⁸O and δ²H, e.g., PC1, PC4 of Cluster I, and PC2 of Cluster II together showed that stable isotope compositions were controlled by reclaimed water (main water source) and strong evaporation.

Nitrogen and phosphorus are the nutrient elements for phytoplankton growth. Consequently, N and P contents would be affected by the biomass of phytoplankton [49]. Therefore, the relationship of nutrients with Chl.a were complicated. Sometimes, a significant and positive correlation in a certain period and a disproportionate relationship both could be found in the water environment [35,81]. There must be a relatively excessive amount of N (for P-restricted water) or P (for N-restricted water). Nutrients would be generally consumed by phytoplankton according to the Redfield ratio [82,83]. This case could be identified by the ratio of N/P. As a result, the surplus nutrient would not contribute to eutrophication. In fresh water, the N/P ratio was less than 7, indicating that N was the possible restrictive nutrient, and if the N/P ratio was greater than 7, then P was the possible restrictive one. Except SY02 in May, the ratio values of all monitoring stations were much higher than 7. Therefore, P was the restricted nutrient. Besides, hydrodynamic conditions such as temperature, light, water volume, and flow rate were also important influencing factors [49,81].

3.4. Assessment of Water Quality

According to the water quality standard for reclaimed water used for a scenic environment (GB/T18921-2019) in China [84], the average pH of reclaimed water (Table 1) was 7.93 ± 0.25, which was in the range of 6–9; the means of NH₃-N, TN, and TP were 0.08 ± 0.03 mg/L, 17.18 ± 4.47 mg/L, and 1.05 ± 0.46 mg/L, respectively. Correspondingly, their threshold values were 5 mg/L, 15 mg/L, and 0.5 mg/L. Therefore, TN and TP of reclaimed water should be reduced to meet the current standard requirement [84]. In terms of the surface water environment quality standard (GB3838-2002) in China [85], Class I is for the source of drinking water, the National Nature Reserve; Class II is for the centralized drinking water surface water source with the first level protected areas, rare aquatic habitat; and Class V is for the agricultural water area and the general landscape requirements of the waters. Average value of TN (9.62 ± 5.29 mg/L) and TP (0.60 ± 0.55 mg/L) in river water exceeded their limited value of the V class (2.0 mg/L, 0.4 mg/L). While, NH₃-N (0.38 ± 0.42) in river water was in the range of the I (0.15 mg/L) and II (0.50 mg/L) class, and DO (6.20 ± 2.50 mg/L) was higher than the threshold value in the II (6.00 mg/L) class.

On the basis of the sodium adsorption ratio (SAR), proposed by Richards [86], levels were divided into four classes (S1: <40; S2: 40–90; S3: 90–150; and S4: >150), and the corresponding harmful extents were low, medium, high, and very high, respectively. SAR of reclaimed water and river water were 57.80 ± 5.43 and 56.84 ± 10.27, respectively. Both were in the range of 46.46 ± 12.66–62.02 ± 6.92. As a result, the level was S2, indicating a medium harmful level. TP = 0.02 mg/L and TN = 0.2 mg/L, which were the internationally recognized threshold for eutrophication [87]. TP and TP in reclaimed water and river water were significantly higher than these threshold values. Chl.a in almost all samples were higher than 40 μg/L, except for three monitoring stations (SY01, SY02, and SY03) and some individual samples (SY04 in March; SY06 and SY07 in July; SY09 in September and November; and SY08 in November), which indicated severe eutrophication [87,88]. Except for SY01 and SY02, the remaining samples were at the eutrophication level, with Chl.a > 7 μg/L.

4. Conclusions

Chemometrics and multivariate statistics were used to study the characteristics and controlling factors in Chaobai River water, replenished by reclaimed water. The main conclusions were as follows. All water was oxidized and alkaline, which was beneficial for nitrification. Nitrate was the main nitrogen form in reclaimed and river water. Depleted and enriched stable isotopes were in reclaimed water and river water, respectively. TN and TP of reclaimed water exceeded the threshold of the reclaimed water reuse standard and Class V in the surface water quality criteria. Most river water was at the severe eutrophication level. The sodium adsorption ratio indicated a medium harmful level for irrigation purposes. Significant spatial and temporal variation was explored by a cluster analysis. Five months are classified into two distinct groups (I: March, May; II: July, September, and November). Nine stations were classified into two clusters (upstream: SY01, SY02, SY03, SY05, and SY06 and downstream: SY04, SY07, SY08, andSY09). Six parameters (Cl⁻, SO₄²⁻, K⁺, Na⁺, Mg²⁺, and TDS) had significant upward temporal variation. Twelve parameters (DO, EC, HCO₃⁻, K⁺, Na⁺, Ca²⁺, TDS, NO₂-N, NO₃-N, TN, TP, and Chl.a.) had a significant downward spatial trend. While, four parameters (Mg²⁺, NH₃-N, δ¹⁸O, and δ²H) were the opposite. The Gibbs plot showed that river water chemistry was mainly controlled by reclaimed water or the interaction of river water with soil/rock. The ionic relationship and principal component analysis showed that river water had undergone the dissolution of carbonate, calcite, and silicate minerals, cation exchange, a process of nitrification, photosynthesis of phytoplankton, and stable isotope enrichment by strong evaporation, and gypsum and salt rock have a potential dissolution process, after reclaimed water was replenished to the river. We think that water quality of reclaimed water should be furtherly improved to avoid river water eutrophication, especially for nitrogen and phosphorus.