Research on Enhancing the Performance of Pre-Treatment Systems for Saline–Alkaline Agricultural Drainage in Southern Xinjiang

Abstract

Freshwater scarcity in southern Xinjiang has intensified the need for effective utilization of saline–alkaline agricultural drainage. This study evaluates pre-treatment technologies for reverse osmosis (RO) systems to improve water quality and mitigate membrane fouling. Three processes were tested: coagulation–sedimentation–media filtration (G₁), micro-flocculation–media filtration (G₂), and micro-flocculation (G₃) combined with ultrafiltration and varying polyaluminum chloride (PAC) dosages (0–15 mg·L⁻¹). Results show that G₁ and G₂ significantly outperform G₃ in removing turbidity, organic matter, and inorganic ions, achieving SDI₁₅ < 5 and turbidity < 0.3 NTU, meeting RO feedwater standards. Optimal performance occurred at the 7.5–10 mg·L⁻¹ coagulant dosage range, effectively controlling flux decline and fouling. The integrated pre-treatment–ultrafiltration system provides a robust technical framework for saline–alkaline water desalination, offering practical guidance for sustainable water resource utilization in arid agricultural regions.

1. Introduction

Water scarcity and pronounced imbalances between supply and demand are major bottlenecks constraining sustainable economic development in arid regions [1]. To address the severe challenges posed by freshwater shortages, seawater desalination technology has gradually become a key technical support for obtaining freshwater resources in arid, semi-arid and coastal regions [2,3]. Large-volume flood irrigation for salt leaching is commonly used in the process of ameliorating saline–alkali land [4,5]. This process generates large quantities of saline–alkali drainage water characterized by high mineralization, high hardness, and excessive organic matter, for which there is currently no systematic strategy for resource utilization [6]. This not only results in resource wastage but also poses risks of secondary salinization and ecological degradation to downstream desert ecosystems. Therefore, the rational development and utilization of the abundant saline–alkali water resources with great exploitation potential in southern Xinjiang [7,8,9] are of great importance for alleviating the water crisis and improving salinized soils [10,11].

Reverse osmosis (RO) technology, owing to its high desalination efficiency, low energy consumption, and convenient operation, has been widely applied in seawater and brackish water desalination [12]. However, membrane fouling and scaling-induced flux decline, shortened membrane lifetime, and increased operating costs pose major challenges to its practical application [13]. As the “front-end barrier” in membrane-based desalination processes, the pre-treatment stage directly determines product water quality, energy efficiency, and the overall operating costs of the system [14]. Efficient pre-treatment is therefore crucial for ensuring the long-term and stable operation of RO systems. Pre-treatment technologies for RO systems mainly comprise conventional processes (such as coagulation, sedimentation, filtration, oxidation, and adsorption) and membrane-based processes (such as microfiltration and ultrafiltration) [15]. Although both types of pre-treatment can effectively mitigate fouling in RO systems, the quality of saline–alkali water varies markedly among regions due to differences in pollution sources [16]. When treating such water sources, single pre-treatment processes often suffer from technical bottlenecks, including excessive chemical consumption, suboptimal filtration performance, and poor resistance to water quality fluctuations, which, if not properly controlled, can readily lead to fouling and clogging in RO systems [17,18]. In contrast, integrated pre-treatment technologies can effectively overcome the limitations of single-unit processes in removing complex contaminants, thereby providing complementary and synergistic benefits [19,20]. Previous studies have shown that the addition of coagulants can enhance the filtration performance of the pre-treatment system [21], improve product water quality, and mitigate fouling in downstream membrane units [22,23]. These improvements, in turn, exert a decisive influence on the operating costs and service life of the desalination system [24,25]. When conventional seawater desalination technologies are applied to the treatment of agricultural saline–alkali drainage, they face the dual challenges of high energy consumption and severe membrane-fouling risks [26,27]. By comparison, membrane-based integrated pre-treatment technologies can significantly alleviate these problems. It has been reported that micro-flocculation–filtration technology, which couples coagulation and filtration units while omitting the sedimentation step, not only reduces capital investment and chemical consumption, but also provides stable, high-quality product water, making it particularly suitable for treating slightly polluted water sources characterized by low temperature, low turbidity, and low color [28].

Based on an analysis of the local agricultural saline–alkali water quality characteristics, this study investigates the effects of different pre-treatment processes and their operational parameters on filtration performance, and elucidates the relationship between the product water quality of the pre-treatment system and membrane fouling. The objective is to identify suitable pre-treatment processes and operational parameters for the desalination of agricultural saline–alkali drainage in southern Xinjiang, thereby providing valuable insights for developing new agricultural water-resource recycling models in arid regions.

2. Materials and Methods

2.1. Test Site and Raw Water Quality

The present study was conducted through field experiments at the 23rd Brigade of the 13th Regiment, First Division of the Xinjiang Production and Construction Corps, located in Alar City, China (40°32′17″ N, 81°31′57 ″E). The water intake point is the primary drainage channel of Alar City, which has an annual basin runoff of approximately 4.73 × 10⁷ m³. Its water quality and discharge volume are relatively stable and therefore considered representative. The characteristics of the raw water are shown in

Table 1. Raw Water Quality Parameters for Agricultural Drainage.

Indicator	Numerical Value	Indicator	Numerical Value
Turbidity/NTU	10.8 ± 1.09	UV254/cm−1	0.096 ± 0.006
EC/(mS·cm−1)	11.68 ± 1.94	Chl/(mg·L−1)	5.17 ± 0.33
Ca2+/(mg·L−1)	308.14 ± 12.28	TOC/(mg·L−1)	35.89 ± 2.83
Mg2+/(mg·L−1)	320.24 ± 23.26	Al3+/(μg·L−1)	286.97 ± 9.19
pH	8.12 ± 0.14	Na+/(mg·L−1)	2137 ± 105.34
Cl−/(mg·L−1)	2619 ± 73.01

Note: The above values are all average values measured during the experiment.

.

2.2. Experimental Design

The present study employed a split-plot design, with pre-treatment process combinations constituting the main plots (G₁: coagulation–sedimentation–media filtration, G₂: micro-flocculation–media filtration, G₃: micro-flocculation), and coagulant dosage serving as the subplots 0, 2.5, 5, 7.5, 10, 12.5, and 15 mg·L⁻¹ (T₁–T₇), resulting in 21 treatment configurations. The permeate water quality obtained under different pre-treatment processes and coagulant dosages was comparatively analyzed to identify the optimal pre-treatment scheme. On this basis, a pre-treatment–ultrafiltration integrated system was constructed to mitigate specific flux decline and suppress membrane fouling and scaling.

Materials: Polyaluminum chloride (PAC, 28% purity) was purchased from Ruihua Water Purification Materials Co., Ltd. (Zhengzhou, Henan, China). AFM active glass filter media was supplied by Daisila Technology Group Co., Ltd. (Guangzhou, China). Quartz sand was obtained from Xinjiang Yihuaipudao Environmental Protection Equipment Co., Ltd. (Urumqi, China). The ultrafiltration membrane was procured from Guangdong Jucheng Bayer Technology Co., Ltd. (Guangzhou, China), with a pore size of 0.1 μm, a molecular weight cut-off of 100 kDa, and an operating pressure of 0.1–0.2 MPa.

The pre-treatment system was designed with a processing capacity of 1 m³/h. The mixing unit employed polyaluminum chloride (28% purity), which was dosed using a peristaltic pump. The hydraulic retention time (HRT) of the inclined-plate sedimentation tank was set to 30 min. The media filtration system adopted a two-stage configuration: the first-stage media filter used quartz sand as the filter medium (particle size: 2–4 mm), while the second-stage filter utilized AFM (Activated Filter Media) glass media (particle size: 0.4–1 mm). The thickness of each filter layer was maintained uniformly at 1.0–1.2 m.

2.3. Testing Metrics and Methods

2.3.1. Determination of Physicochemical Properties of Water

Turbidity in the water samples was measured using a WZS-172 turbidity meter (Leici, Shanghai, China). The silt density index (SDI₁₅) was determined with an SDI tester (QZDY Pollution Index Tester, Shanghai, China). The concentration of Cl⁻ ions was analyzed using a fully automated chemical analyzer (SmartChem 200, Westco Scientific Instruments Inc., Brookfield, CT, USA), while the concentration of Na⁺ ions was measured using a flame photometer (FP6410, Shanghai Precision Instruments Co., Ltd., Shanghai, China). The concentrations of Ca²⁺ and Mg²⁺ ions were determined using the ethylenediaminetetraacetic acid (EDTA) titration method.

2.3.2. Determination of Organic Matter Indicators

Chlorophyll content in the water samples was determined in strict accordance with the technical specifications of Measurement of Phytoplankton in Water Quality—Filter Membrane-Spectrophotometric Method (HJ 1215-2021) [29]. During the experiment, water samples were periodically collected at each sampling point, filtered through a membrane using 1 L of sample water, rapidly frozen with liquid nitrogen, and subsequently stored at −80 °C in an ultra-low-temperature freezer. Quantitative analysis was conducted by measuring the absorbance of particulate matter retained on the membrane. Before testing, the filter membrane was soaked in methanol for 4 h to extract non-algal pigments, and a blank membrane was used as the reference. The phytoplankton content was then calculated based on the absorbance difference. Total organic carbon (TOC) was determined using high-temperature catalytic oxidation with nondispersive infrared detection (Shimadzu TOC-L; ISO 8245:1999) [30]. Ultraviolet absorbance (UV₂₅₄) was measured using a UV-visible spectrophotometer at a wavelength of 254 nm.

2.3.3. Metal Ion Determination

The water samples were acidified with 1% HNO₃ and stored in polypropylene bottles. Iron and manganese ions were determined by flame atomic absorption spectrometry (FAAS), with Fe analyzed at 248.3 nm and Mn at 279.5 nm. A 0.1% LaCl₃ solution was added to eliminate interference from coexisting ions. Aluminum ions were determined by graphite furnace atomic absorption spectrometry (GFAAS) using a pyrolytic-coated graphite tube. A 0.5% Mg(NO₃)₂ matrix modifier was added, with the temperature program set as 110 °C (drying), 1300 °C (ash), and 2500 °C (atomization). The calibration curves for each element exhibited a linear correlation coefficient (R²) greater than 0.995.

2.3.4. Membrane Flux

The ultrafiltration membrane was soaked in ultrapure water for at least 24 h, during which the water was replaced 2–3 times to remove the protective coating on the membrane surface. Before conducting the filtration test, the membrane was pre-filtered with ultrapure water, and the experiment was initiated once the system flux had stabilized. During the experiment, an electronic balance was used to record the real-time weight of the permeate from the ultrafiltration system, which was then used to calculate the membrane flux ratio, defined as J/J₀.

J₀ is the membrane flux during ultrapure water filtration; J is the membrane flux when the water produced from medium filtration is used as the water source.

Membrane flux calculation formula:

$$J={\displaystyle \frac{V}{At}}$$

In the formula, J is the membrane flux (m³·m⁻²·h⁻¹); V is the filtration volume of the ultrafiltration system (m³); A is the membrane area (m²); and t is the filtration time (h).

2.4. Data Processing and Analysis

Data processing was conducted using Microsoft Excel 2019. 8. Analysis of variance (ANOVA) and significance testing were conducted with SPSS 26 (IBM Corp., Chicago, IL, USA). Origin 2021 and Microsoft PowerPoint 2019 were employed for graph plotting. The entropy weight-TOPSIS multi-objective decision analysis method was employed to evaluate the optimal solution for pre-treatment processes and dosing amounts.

3. Results and Analysis

3.1. Comparative Analysis of the Optimal Flocculation Conditions for the Pre-Treatment Unit

3.1.1. Removal Performance of the Pre-Treatment Unit for Particulate Matter in Water

The removal efficiency of turbidity in water under different coagulant dosage conditions for each pre-treatment process is shown in Figure 1. In the context of the G₁ and G₂ processes, an increase in chemical dosage resulted in a reduction in the turbidity of the system’s treated water. The turbidity removal rates at T₂–T₆ exhibited an enhancement of 4.41–13.24% in comparison to T₁. However, the G₃ pre-treatment process proved ineffective in achieving adequate turbidity removal, and the application of higher chemical dosages resulted in an increase in the turbidity of the treated water. At equivalent chemical dosages, the G₁ and G₂ processes exhibited superior particulate removal performance in comparison to G₃, achieving turbidity removal rates that were 44.89% and 46.37% higher than those of G₃, respectively. The interaction of pre-treatment processes and dosage was found to be a significant factor in the performance of G₁T₇ treatment, which exhibited optimal particulate removal performance, showing improvements of 13.23% to 34.98% over CK (control without coagulant addition).

3.1.2. Removal Performance of the Pre-Treatment Unit for Organic Matter in Water

The removal effects of the pre-treatment units on chlorophyll (Chl), TOC, and UV₂₅₄ in water under different coagulant dosages are shown in Figure 2. The results show that the G₁ and G₂ pre-treatment units performed better than G₃ in removing Chl, TOC, and UV₂₅₄, with removal rates improving by 16.39% and 21.78% (Chl), 19.62% and 24.89% (TOC), and 21.71% and 24.19% (UV₂₅₄), respectively. The Chl removal rate increased monotonically with the coagulant dosage, achieving optimal removal at 15 mg·L⁻¹. However, the removal effects of UV₂₅₄ and TOC exhibited threshold values, with the best removal effects occurring at 10 mg·L⁻¹ (T₅) and 12.5 mg·L⁻¹ (T₆), respectively.

3.2. Filtration Effectiveness and Membrane Fouling Process of the Integrated Pre-Treatment

3.2.1. Removal Efficiency of Turbidity and Organic Matter

The filtration effect of the integrated pre-treatment units on water pollutants is shown in

Table 2. Filtration Performance of Contaminants Using Ultrafiltration Technology Coupled with Various pre-treatment Systems.

Treatments		NTU	TOC	UV254	Chl
G1-UF	T1	0.24 ± 0.031 a	17.19 ± 0.27 f	0.060 ± 0.003 d	0.50 ± 0.016 e
	T2	0.26 ± 0.021 a	16.22 ± 0.16 e	0.052 ± 0.005 cd	0.43 ± 0.006 d
	T3	0.24 ± 0.035 a	14.69 ± 0.59 d	0.048 ± 0.007 bc	0.40 ± 0.004 c
	T4	0.24 ± 0.026 a	13.29 ± 0.14 bc	0.039 ± 0.008 ab	0.39 ± 0.006 c
	T5	0.25 ± 0.035 a	12.48 ± 0.37 a	0.034 ± 0.008 a	0.35 ± 0.008 b
	T6	0.25 ± 0.036 a	12.81 ± 0.67 ab	0.037 ± 0.006 ab	0.31 ± 0.010 a
	T7	0.24 ± 0.035 a	13.55 ± 0.15 c	0.038 ± 0.008 ab	0.31 ± 0.008 a
G2-UF	T1	0.24 ± 0.031 a	19.29 ± 0.13 d	0.060 ± 0.009 c	0.53 ± 0.016 d
	T2	0.27 ± 0.031 a	19.17 ± 0.55 d	0.053 ± 0.007 bc	0.45 ± 0.008 c
	T3	0.24 ± 0.030 a	17.62 ± 0.59 c	0.049 ± 0.007 abc	0.41 ± 0.004 b
	T4	0.27 ± 0.026 a	17.23 ± 0.60 e	0.040 ± 0.009 ab	0.43 ± 0.031 b
	T5	0.28 ± 0.015 a	15.25 ± 0.30 a	0.035 ± 0.007 a	0.37 ± 0.004 a
	T6	0.28 ± 0.006 a	15.52 ± 0.19 a	0.038 ± 0.008 ab	0.35 ± 0.009 a
	T7	0.27 ± 0.020 a	16.42 ± 0.27 b	0.039 ± 0.008 ab	0.35 ± 0.005 a
G2-UF	T1	0.26 ± 0.026 a	19.98 ± 0.21 d	0.064 ± 0.007 c	0.80 ± 0.008 e
	T2	0.25 ± 0.012 a	19.27 ± 0.26 bcd	0.063 ± 0.006 c	0.63 ± 0.015 d
	T3	0.26 ± 0.015 a	19.74 ± 0.48 cd	0.059 ± 0.005 c	0.56 ± 0.008 c
	T4	0.27 ± 0.006 a	18.72 ± 0.71 b	0.057 ± 0.004 bc	0.52 ± 0.013 b
	T5	0.26 ± 0.015 a	17.63 ± 0.57 a	0.045 ± 0.004 a	0.51 ± 0.006 b
	T6	0.27 ± 0.012 a	17.36 ± 0.13 a	0.046 ± 0.005 a	0.51 ± 0.004 b
	T7	0.27 ± 0.020 a	18.88 ± 0.76 bc	0.048 ± 0.007 ab	0.49 ± 0.004 a
G		*	**	**	**
T		ns	**	**	**
G × T		ns	**	ns	**

Note: Different lowercase letters in the table indicate statistically significant differences (p < 0.05). G denotes the pre-treatment process, T represents the coagulant dosage, and G × T indicates the interaction between the two factors. Symbols * and ** denote significance at the 0.05 and 0.01 levels, respectively, while ns indicates no significant difference.

. A significant interaction between the integrated pre-treatment units and coagulant dosage was observed for the removal of turbidity and organic matter. The study showed that the permeate water quality was substantially improved following integrated pre-treatment, with turbidity removal rates exceeding 97.22% and permeate turbidity maintained below 0.3 NTU. These values meet the feed-water requirements of reverse osmosis systems (turbidity < 1 NTU, preferably <0.3 NTU). Compared with G₃-UF, the G₁-UF and G₂-UF pre-treatment processes demonstrated enhanced removal efficiencies for TOC, UV₂₅₄, and chlorophyll, with increases of 3.10% and 12.48% (TOC), 10.11% and 10.91% (UV₂₅₄), and 3.14% and 3.66% (Chl), respectively. Furthermore, increasing the coagulant dosage significantly improved the removal of organic matter. For the T₂–T₆ treatments, the removal rates were higher than those of T1 by 1.67–10.31% (TOC), 5.56–24.31% (UV₂₅₄), and 2.06–4.38% (Chl).

3.2.2. Effects of Specific Flux and SDI₁₅ on Ultrafiltration Membrane Fouling

The study found that the raw water and the permeates from the G₁, G₂, and G₃ pre-treatment units all exceeded the measurement range of the SDI₁₅ test, indicating that they failed to meet the reverse osmosis feed-water requirement (SDI₁₅ < 5). As shown in Figure 3, coupling the pre-treatment processes with ultrafiltration significantly reduced the SDI₁₅ values of the permeates. Notably, the SDI₁₅ values for the G₁-UF and G₂-UF integrated pre-treatments were both below 5, with reductions of 9.76% and 12.34%, respectively, compared with the G₃-UF process. With increasing coagulant dosage, the SDI₁₅ removal performance of each integrated pre-treatment unit exhibited a nonlinear pattern, initially declining and subsequently improving. Among all treatment combinations, G₁T₄-UF achieved the greatest reduction in SDI₁₅, with a decrease of 2.7 in the permeate.

The variation in membrane specific flux decay for each integrated pre-treatment unit under increasing coagulant dosage is shown in Figure 4. As the system operating time increased, the application of conventional hydraulic backwashing partially restored the membrane specific flux; however, an overall declining trend remained evident across all integrated pre-treatment units. Notably, specific flux decay in the G₁-UF and G₂-UF units was more effectively controlled compared with the G₃-UF unit. In addition, as the coagulant dosage increased, the specific flux decay of each integrated pre-treatment unit first decreased and then increased. Among all treatments, G₁T₄ and G₂T₄ exhibited the lowest specific flux decay values, reaching 0.81, and all pre-treatment processes showed better control over membrane flux decay rate after coagulant addition compared to CK (no coagulant added).

3.2.3. Removal Effectiveness of the Integrated Pre-Treatment Unit for Ions in Water

The removal effectiveness of each integrated pre-treatment unit for inorganic salt ions in water is shown in

Table 3. Water Quality Parameters for Produced Water from Each Pre-treatment System.

Treatments		Ca2+	Mg2+	Al3+	EC
G1-UF	T1	266.81 ± 7.25 d	285.50 ± 4.02 d	133.97 ± 5.04 d	11.28 ± 0.07 b
	T2	253.69 ± 5.44 c	273.93 ± 6.40 c	141.40 ± 6.51 d	11.20 ± 0.01 c
	T3	238.24 ± 6.97 b	258.46 ± 3.20 b	95.88 ± 5.20 c	11.12 ± 0.03 a
	T4	222.48 ± 4.74 a	236.60 ± 5.60 a	72.68 ± 4.99 b	11.06 ± 0.04 a
	T5	225.75 ± 5.48 a	234.90 ± 4.06 a	52.02 ± 7.04 a	10.94 ± 0.06 a
	T6	225.86 ± 5.00 a	232.60 ± 7.22 a	71.31 ± 6.60 b	11.08 ± 0.05 a
	T7	224.71 ± 6.36 a	230.92 ± 2.31 a	93.49 ± 5.05 c	11.12 ± 0.04 a
G2-UF	T1	275.96 ± 8.56 a	291.08 ± 2.59 d	133.25 ± 5.69 e	11.32 ± 0.02 d
	T2	252.79 ± 6.00 b	277.94 ± 6.12 c	141.48 ± 5.96 e	11.22 ± 0.03 c
	T3	246.49 ± 5.20 b	265.73 ± 5.61 b	105.95 ± 5.86 d	11.18 ± 0.04 bc
	T4	230.42 ± 5.85 a	242.22 ± 1.00 a	74.88 ± 4.70 b	11.16 ± 0.05 bc
	T5	233.11 ± 5.15 a	241.87 ± 1.40 a	55.89 ± 4.63 a	11.07 ± 0.04 a
	T6	232.02 ± 7.81 a	241.22 ± 0.84 a	73.82 ± 4.19 b	11.09 ± 0.02 a
	T7	231.00 ± 8.59 a	239.82 ± 1.82 a	95.15 ± 4.64 c	11.12 ± 0.02 ab
G3-UF	T1	291.65 ± 7.64 c	300.26 ± 2.19 c	133.04 ± 5.98 d	11.53 ± 0.01 e
	T2	281.88 ± 6.55 b	293.27 ± 5.24 bc	146.73 ± 7.78 e	11.44 ± 0.03 d
	T3	276.38 ± 2.26 ab	284.90 ± 7.43 a	124.80 ± 7.70 d	11.35 ± 0.04 c
	T4	275.13 ± 1.98 ab	281.72 ± 2.81 a	91.54 ± 4.18 b	11.24 ± 0.08 b
	T5	272.77 ± 1.57 a	281.91 ± 5.77 a	66.51 ± 6.02 a	11.15 ± 0.06 a
	T6	271.67 ± 1.66 a	286.29 ± 0.95 ab	81.90 ± 7.11 b	11.23 ± 0.02 b
	T7	275.11 ± 3.55 ab	287.74 ± 1.52 ab	107.73 ± 7.57 c	11.35 ± 0.04 c
G		**	**	**	**
T		**	**	**	**
G × T		**	**	*	*

Note: Different lowercase letters in the table indicate statistically significant differences (p < 0.05). G denotes the pre-treatment process, T represents the coagulant dosage, and G × T indicates the interaction between the two factors. Symbols * and ** denote significance at the 0.05 and 0.01 levels, respectively, while ns indicates no significant difference.

. The results indicate that both the type of integrated pre-treatment unit and the coagulant dosage significantly influenced the permeate water quality, with a notable interaction observed between the two factors. As shown in Table 3, the G₁-UF and G₂-UF pre-treatment units exhibited better removal performance for inorganic pollutants in water than G₃-UF, with improvements of 11.26–13.31% (Ca²⁺), 9.64–11.74% (Mg²⁺), 3.57–4.55% (Al³⁺), and 1.38–1.82% (EC), respectively. Furthermore, increasing the coagulant dosage further enhanced the filtration performance of each pre-treatment system. For the T₂–T₆ treatments, the removal rates of inorganic salt ions were higher than those of T1 by 4.98–11.51% (Ca²⁺), 3.30–12.32% (Mg²⁺), and 3.41–8.38% (EC).

3.3. Comprehensive Evaluation Based on Entropy Weight-TOPSIS Method

The present study employed the entropy-weighted TOPSIS method for comprehensive evaluation in order to determine the effects of different pre-treatment processes on turbidity, SDI₁₅, membrane specific flux, organic pollutants, and inorganic pollutants (

Table 4. TOPSIS evaluation of pre-treatment filtration performance (turbidity, SDI₁₅, membrane flux and organic pollutants).

Treatment		Euclidean Distances		Comprehensive Score Index	TOPSIS Rank
		d+	d−
G1-UF	T1	0.25	0.20	0.45	12
	T2	0.23	0.23	0.50	11
	T3	0.13	0.29	0.70	9
	T4	0.07	0.35	0.84	3
	T5	0.02	0.40	0.95	1
	T6	0.04	0.38	0.90	2
	T7	0.09	0.34	0.78	5
G2-UF	T1	0.26	0.18	0.41	14
	T2	0.26	0.20	0.44	13
	T3	0.18	0.24	0.58	10
	T4	0.12	0.30	0.71	8
	T5	0.07	0.36	0.84	3
	T6	0.07	0.35	0.83	4
	T7	0.10	0.32	0.76	7
G3-UF	T1	0.37	0.07	0.17	21
	T2	0.36	0.08	0.17	20
	T3	0.34	0.10	0.23	19
	T4	0.31	0.14	0.31	18
	T5	0.31	0.19	0.38	15
	T6	0.30	0.17	0.35	16
	T7	0.31	0.15	0.32	17

). The evaluation encompassed target weighting, Euclidean distance, comprehensive score calculation, and TOPSIS ranking of varying coagulant dosage combinations across pre-treatment systems. Using the entropy-weighting approach, the relative weights of the indicators were determined as follows: UV₂₅₄ (13.39%), TOC (23.41%), chlorophyll (11.59%), SDI₁₅ (6.47%), Turbidity (16.08%), Al³⁺ (20.16%), and membrane specific flux (8.90%). Based on the comprehensive TOPSIS scores, the top three performing systems were G₁T₅, G₁T₆, and G₁T₄.

4. Discussion

Agricultural saline–alkaline drainage differs fundamentally from typical seawater or industrial wastewaters: it is characterized by high mineralization and salt accumulation and is frequently subject to variable pollutant loads-such as residual agrochemicals and episodic algal blooms. These conditions engender complex interactions among colloids, dissolved electrolytes and organic matter, undermining the ability of conventional pretreatment to simultaneously remove particulate colloids and control dissolved organics. In practice, this complexity accelerates membrane fouling, increases cleaning frequency, and substantially raises energy consumption, thereby constituting a major technical barrier to the wider deployment of desalination for this water source. Selecting an appropriate pre-treatment process is crucial for ensuring the long-term, stable operation of reverse osmosis systems and directly influences the service life of reverse osmosis membranes [31,32,33]. The findings of this study indicate that, compared with the G₃ treatment, the G₁ and G₂ pre-treatment processes exhibit superior removal efficiencies for particulate matter and organic pollutants. This suggests that coagulation alone is insufficient for effectively filtering and intercepting contaminants in saline–alkaline water and further confirms that the incorporation of a media filtration unit can substantially enhance the removal of particulate matter and algal-derived organic substances. These results are consistent with the findings reported by Namazi et al. [34]. Several studies have similarly demonstrated that the addition of coagulants is essential for improving the filtration efficiency of pre-treatment systems [35,36,37]. as coagulants promote electro neutralization of colloidal particles in water, thereby facilitating floc formation and enhancing the removal performance of media filtration [38]. However, the experimental results of this study clearly show that conventional pre-treatment processes (G₁, G₂, G₃) are inadequate for effectively removing SDI₁₅-forming particles from saline–alkaline agricultural drainage in southern Xinjiang. As a result, the permeate water quality fails to meet the feed-water requirements for reverse osmosis systems (SDI₁₅ < 5) [39].

Membrane separation technology employs semi-permeable membranes to selectively permit the passage of specific substances, thereby achieving efficient separation of salts and impurities. Previous studies have reported that integrated technologies can effectively overcome the technical bottlenecks of membrane fouling and clogging associated with the limited filtration efficiency and poor resistance to water-quality fluctuations observed in traditional pre-treatment processes [40]. In this study, the integration of the three pre-treatment processes (G₁, G₂, G₃) with ultrafiltration significantly improved system performance, enabling both permeate turbidity and SDI₁₅ to meet the feed-water requirements for reverse osmosis. This improvement is primarily attributed to the ultrafiltration membrane’s high-efficiency retention of suspended solids, bacteria, and macromolecular organic matter [41], which substantially enhances the feed-water quality for subsequent reverse osmosis treatment [42]. The integrated process demonstrated strong adaptability for treating the specific saline–alkaline water sources investigated in this study. Through synergistic coupling effects, it improved permeate water quality while effectively addressing the limitations of filtration efficiency in conventional pre-treatment processes and mitigating the issues of fouling and clogging in single membrane separation units. Furthermore, the experimental data revealed a significant correlation between the removal efficiency of particulate matter and organic matter in the pre-treatment units and both SDI₁₅ and the membrane specific flux decay rate, consistent with the findings of Zhao et al. [28].

The causes of membrane fouling and clogging are closely associated with the presence of inorganic ions such as calcium and magnesium carbonates and metal hydroxides in water [43]. In coagulation-sedimentation units, colloidal calcium and magnesium are primarily removed through chemical precipitation (CaCO₃/Mg(OH)₂ crystallization) and co-precipitation. In contrast, ultrafiltration relies on membrane pores to retain these precipitated colloids but cannot effectively remove dissolved ions. In this study, it was observed that increasing the coagulant dosage led to elevated concentrations of calcium, magnesium, and aluminum ions in the water. When the coagulant dosage was below 10 mg·L⁻¹, no significant difference in aluminum concentration was observed among the permeates of the G₁-UF, G₂-UF, and G₃-UF systems. However, when the dosage exceeded 10 mg·L⁻¹, aluminum concentrations increased markedly, indicating that excessive coagulant remained unreacted in the water [28]. Therefore, when applying coagulants, it is essential to account for the threshold dosage range. Moreover, the three pre-treatment processes exhibit limited removal of calcium and magnesium ions. To effectively mitigate membrane flux decline and the increased energy consumption associated with inorganic scaling (e.g., Ca²⁺ and Mg²⁺ deposition on the membrane surface), the appropriate addition of antiscalants prior to reverse osmosis feedwater is recommended [44], supplemented by routine chemical cleaning and maintenance to jointly reduce the risk of membrane fouling. Optimizing both the pre-treatment process and the coagulant dosage is an effective strategy for enhancing system retention efficiency and minimizing the fouling and clogging risks in membrane systems. Previous research has demonstrated that the filtration performance of pre-treatment systems does not exhibit a simple, linear positive correlation with coagulant dosage [45]. In this study, the pre-treatment schemes in G₁–G₃ all exhibited superior membrane-fouling control under the coagulant dosage condition T₅. This observation can be explained by a “critical-threshold” effect in coagulation–ultrafiltration systems. At an optimal dosage near T₅, the coagulant promotes aggregation of colloidal particles through charge neutralization and bridging/sweep flocculation, producing large, mechanically robust flocs that are readily retained by the ultrafiltration membrane. Such aggregation substantially reduces pore blocking and cake-layer resistance, thereby maintaining higher membrane flux and a lower flux-decline rate. By contrast, dosages that exceed this threshold can be detrimental: excessive dosing may cause over-neutralization or charge reversal, leading to colloid re-stabilization or formation of fragile flocs that readily break apart; additionally, surplus coagulant can accumulate on the membrane surface to form an adhesive layer or increase stickiness, ultimately accelerating membrane fouling. This indicates that a clear dosage threshold exists for coagulants in the pre-treatment system. Exceeding the optimal dosage range leads to residual coagulant remaining in the water, which then becomes an additional source of contamination and exacerbates membrane fouling [46]. Therefore, optimizing the coagulant dosage is essential for controlling membrane-fouling rates and reducing system operating costs. In particular, precise dosage control is critical for preventing irreversible fouling of reverse osmosis membranes caused by residual aluminum ions. In this study, no significant differences in aluminum concentrations were observed among the permeates of the pre-treatment systems at low to moderate coagulant dosages (<10 mg·L⁻¹). However, at higher dosages (10–15 mg·L⁻¹), aluminum concentrations increased significantly, likely due to the incomplete reaction and subsequent persistence of excess coagulant in the water [28]. These findings underscore the need for appropriate pre-treatment configuration and strict coagulant dosage control to ensure effective filtration performance and prevent contamination of the membrane system. By reducing both membrane-fouling rates and coagulant consumption, the overall operating costs of the system can be further minimized. In summary, practical engineering applications should determine the pre-treatment scheme and optimal coagulant dosage strictly based on the specific water-quality characteristics of the source water.

The operating parameters of the pre-treatment unit are critical for ensuring the safe, stable, and long-term operation of the ultrafiltration system. In this study, periodic hydraulic backwashing effectively restored the flux of the ultrafiltration membrane, indicating that regular backwashing is an essential measure for maintaining flux recovery. However, the flux recovery rate exhibited a gradual decline over time. This decline can be attributed to the formation of a cake layer on the membrane surface by certain pollutants, while other contaminants cause irreversible pore blockage within the membrane structure [47]. Research has shown that membrane pore clogging is the fundamental cause of irreversible flux decline. Although hydraulic backwashing can effectively remove the cake layer on the membrane surface, it has limited impact on pore blockages that have already formed [48]. In practical applications, periodic chemical cleaning (acid and alkali washing) and air–water mixed scrubbing are commonly implemented as enhanced cleaning strategies to restore ultrafiltration membrane flux and maintain permeate water quality and operational stability, thereby addressing issues such as pore clogging and flux decline. To determine the most suitable pre-treatment processes and operational parameters for reverse osmosis desalination of farmland saline–alkali drainage in southern Xinjiang, this study employed the entropy-weighted TOPSIS method to comprehensively evaluate permeate water quality, inorganic and organic pollutant removal, and membrane specific flux decay. Among all configurations, the coagulation–sedimentation–media filtration–ultrafiltration process demonstrated the best overall performance, followed by the micro-flocculation–media filtration–ultrafiltration process. Under the conditions of achieving high permeate quality (SDI₁₅ < 3, Turbidity < 0.1 NTU) and effectively controlling membrane fouling, the micro-flocculation–media filtration process offers significant advantages. By eliminating the sedimentation unit, it reduces construction costs, lowers energy consumption, and minimizes land requirements. Compared with coagulation-sedimentation-media filtration and micro-flocculation alone, the micro-flocculation-media filtration-ultrafiltration process provides superior energy efficiency and overall treatment performance for saline–alkali agricultural drainage in this region. With precise control of coagulant dosage, this integrated process can serve as an effective pre-treatment strategy for saline–alkali drainage desalination in southern Xinjiang, substantially improving filtration performance and suppressing membrane fouling. Following further validation and optimization through pilot-scale studies, it is expected to become a competitive and practical pre-treatment solution for large-scale agricultural saline–alkali drainage desalination.

5. Conclusions

This study examines the selection of effective pre-treatment processes and strategies for controlling membrane fouling in the desalination of saline–alkaline agricultural drainage in southern Xinjiang. The experimental results demonstrate that incorporating flocculation and media filtration units significantly improves permeate water quality and reduces the risk of ultrafiltration membrane fouling. Moreover, precise coagulant dosing is critical for optimizing system performance, as it directly influences filtration efficiency and fouling mitigation. This study emphasizes that accurate dosage control is essential in practical applications to ensure stable and efficient system operation. The findings obtained in this work confirm the feasibility and effectiveness of the proposed integrated pre-treatment approaches for enhancing desalination performance in saline–alkaline farmland drainage. These results provide theoretical and technical support for developing robust pre-treatment strategies applicable to agricultural water reuse in arid regions.