Halophyte-Mediated Metal Immobilization and Divergent Enrichment in Arid Degraded Soils: Mechanisms and Remediation Framework for the Tarim Basin, China

Abstract

Understanding heavy metal behavior in arid saline soils is critical for phytoremediation in degraded lands. This study investigated metal distribution and plant enrichment in the Tarim Basin using 323 soil and 55 plant samples (Populus euphratica, Tamarix ramosissima, cotton, jujube). Analyses included redundancy analysis (RDA) and bioconcentration factor (BCF) assessments. Key findings reveal that elevated salinity (total salts, TS > 200 g/kg) and alkalinity (pH > 8.5) immobilized As, Cd, Cu, and Zn. Precipitation and competitive leaching reduced metal mobility by 42–68%. Plant enrichment strategies diverged significantly: P. euphratica hyperaccumulated Cd (BCF = 1.59) and Zn (BCF = 2.41), while T. ramosissima accumulated As and Pb (BCF > 0.05). Conversely, cotton posed Hg transfer risks (BCF = 2.15), and jujube approached Cd safety thresholds in phosphorus-rich soils. RDA indicated that pH and total salinity (TS) jointly suppressed metal bioavailability, explaining 57.6% of variance. Total phosphorus (TP) and soil organic carbon (SOC) enhanced metal availability (36.8% variance), with notable TP-Cd synergy (Pearson’s r = 0.42). We propose a dual-threshold management framework: (1) leveraging salinity–alkalinity suppression (TS > 200 g/kg + pH > 8.5) for natural immobilization; and (2) implementing TP control (TP > 0.8 g/kg) to mitigate crop Cd risks. P. euphratica demonstrates targeted phytoremediation potential for degraded saline agricultural systems. This framework guides practical management by spatially delineating zones for natural immobilization versus targeted remediation (e.g., P. euphratica planting in Cd/Zn hotspots) and implementing phosphorus control in high-risk croplands.

1. Introduction

Heavy metal pollution has become a global environmental issue, posing a potential threat to ecological security, especially in arid regions. Authoritative international studies indicate that approximately 25% of agricultural soils worldwide are contaminated by varying degrees of heavy metals [1]. This issue is particularly acute in arid regions, where the interplay between heavy metals and salinity creates unique environmental dynamics [2]. Arid ecosystems exhibit distinctive hydro-saline dynamics: intense evaporation, limited precipitation, and fluctuating groundwater tables drive extreme salt accumulation (e.g., NaCl, CaSO₄) and alkalization (pH > 8.5), fundamentally altering metal speciation and mobility. These conditions promote salt crust formation, enhance metal adsorption to clay-organic complexes, and trigger competitive ion exchange (e.g., Na⁺ displacing Cd²⁺), collectively reducing metal bioavailability—a phenomenon less pronounced in humid saline soils [3]. China’s “Soil Pollution Prevention and Control Action Plan” designates heavy metal pollution control as a key task, while the “14th Five-Year Plan for National Ecological Protection” specifically emphasizes strengthening research on heavy metal migration patterns under the unique environmental conditions of the arid northwest.

Extensive literature has reported on the spatial distribution, ecological risks, and migration mechanisms of heavy metal pollution in soils, with numerous case studies and mechanistic insights having been accumulated in typical arid regions such as the Hetao Plain and the oasis agricultural areas of Xinjiang. For instance, in the Hetao Irrigation District, systematic sampling and geostatistical analyses have revealed significant variability in elements such as As and Cd, closely linked to irrigation practices. In the Tarim River Basin, preliminary studies have identified enrichment tendencies of Pb and Zn in Populus euphratica and Tamarix ramosissima communities [4]. Globally, halophytes such as Salicornia europaea and Suaeda salsa have been successfully utilized for phytoremediation in saline soils, demonstrating significant potential for metal accumulation and soil restoration. Methodologically, geostatistics, GIS-based spatial interpolation, single-factor pollution indices, and the Nemerow composite index have been widely employed to support regional risk classification. Laboratory simulation experiments have further examined the influence of saline–alkaline stress on heavy metal speciation, providing a foundation for understanding metallic behavior under alkaline conditions [5]. Notably, emerging research directions such as microbe-plant joint remediation, adaptive mechanisms of halophytes, and nanoparticle-assisted remediation have shown significant progress, offering novel perspectives for pollution mitigation in arid zones [6]. Nevertheless, notable limitations persist in understanding the behavior of heavy metals under saline–alkaline conditions in arid regions. Many studies focus on individual heavy metals; for example, Irfan et al. (2022) investigated only the migration of Cd in the Hetao Irrigation District without addressing multi-metal interactions or ion-exchange effects [7]. Moreover, research remains largely confined to macro-scale statistics or idealized controlled experiments, lacking systematic field observations under realistic multi-pollutant scenarios [8]. Regarding plant responses, while some studies have addressed Cd tolerance in Tamarix [9] or salt stress responses in cotton, comparative analyses of typical halophytes (e.g., Populus euphratica, Tamarix ramosissima) and economic crops (e.g., jujube, cotton) within the same contaminated habitats are scarce [4]. This gap impedes a mechanistic understanding of root-soil interface interactions between salts and metals, as well as the principal factors driving differential metal uptake strategies in plants. Furthermore, existing studies heavily rely on total concentration analyses and empirical modeling, with insufficient mechanistic exploration of micro-interfacial chemical processes, key biogeochemical pathways, and plant root responses under specific ion stress. These limitations restrict the development of process-based models applicable to saline–alkaline environments and hinder effective remediation practices [6].

This study focuses on the Tarim River Basin, with the aim of investigating the migration and accumulation of multiple heavy metals (e.g., Cd, Pb) in the soil–plant system under high-salinity conditions in arid regions. A multidisciplinary approach was adopted, incorporating a systematic “grid + patch” sampling design across 323 soil sites and 55 plant sites. In accordance with the Specification for Land Quality Geochemical Assessment, 14 soil parameters and 8 heavy metal indicators in plants were measured. The research aims to elucidate the interactive mechanisms among “salinity–heavy metals–plants” under high-salinity stress, clarify the differential uptake strategies of various plant types (economic crops versus halophytes), and ultimately develop a theoretical model for heavy metal migration and accumulation applicable to arid environments. The findings will provide core scientific support and a decision-making basis for regional pollution risk management, agricultural product safety, and ecological restoration initiatives.

2. Materials and Methods

2.1. Study Area

The study area is situated on the northeastern margin of the Taklimakan Desert, Xinjiang (Figure 1). It spans geographical coordinates of 86°59′–88°02′ E and 40°55′–41°09′ N, covering an area of approximately 1875 km². Bounded by the Taklimakan Desert to the west and the Kuruktag Desert to the east, the region represents a typical arid-zone oasis. It experiences significant diurnal temperature fluctuations, low precipitation, and high evaporation rates, characteristic of an arid desert climate. Dominant soil types include Oasis Orthic Anthrosols, Aeolian Sandy Soils, Luvisols, and Gray Forest Soils, all derived from Quaternary sediments with no bedrock outcrops. The soil matrix primarily comprises fine sand and silt-fine sand, exhibiting well-developed salinization. Natural vegetation is dominated by T. ramosissima, P. euphratica, Apocynum venetum, and Pulicaria dysenterica [10]. The Tarim River traverses the area, serving as the primary water source for local production and domestic use. Leveraging unique resources in soil-water, light-heat, geography, and biodiversity, the region has established distinctive industrial clusters centered on Gossypium hirsutum (cotton), Ziziphus jujuba (jujube), Prunus armeniaca, Humulus lupulus, and Cervus elaphus farming. A commercial-scale, branded industrial layout has begun to take shape.

2.2. Sample Collection and Testing

In accordance with the Specification for Land Quality Geochemical Assessment (DZ/T 0295-2016) [11], this study implemented a stratified “grid + patch” sampling protocol. A total of 323 surface soil specimens (0–30 cm depth) were collected, targeting the primary anthropogenic metal accumulation zone (> 90% deposited heavy metals) and the core rhizosphere activity horizon (containing > 75% fine roots of annual crops/shrubs). Vegetation sampling strategy prioritized functionally dominant species: (i) Native halophytes: T. ramosissima (salt-secreting shrub) and P. euphratica (deep-rooted phreatophyte), collectively representing > 85% of natural riparian cover; (ii) Economically significant crops: cotton and jujuba, occupying > 60% of agricultural land area. This design facilitates comparative assessment of ecological adaptation versus food chain transfer risks. Spatially stratified sampling ensured vegetation representativeness: T. ramosissima (14 sites): Distributed across riparian cores (> 45% vegetation cover) with 500 m grid spacing to resolve salinity gradients (TS: 50–380 g/kg). P. euphratica (14 sites): Positioned along groundwater flow paths (> 40% woody cover, 3–8 m root depth). Cotton (13 sites): Located in 12 contiguous irrigation districts (68% farmland coverage), density-scaled to cultivation intensity. Jujube (14 sites): Sampled from mature orchards (>10-year cultivation, 73% regional coverage) with documented phosphorus management history. Plant samples were collected at the mature growth stage (for P. euphratica and T. ramosissima: fully developed adult trees; for cotton: boll opening stage; for jujube: fruit ripening stage). The root systems were carefully excavated, washed with deionized water to remove adhering soil, and oven-dried at 70 °C to constant weight for heavy metal analysis.

To investigate the spatial distribution characteristics and influencing factors of soil heavy metals in the study area, a total of 14 parameters were determined for the soil samples, including arsenic (As), chromium (Cr), nickel (Ni), copper (Cu), zinc (Zn), mercury (Hg), lead (Pb), cadmium (Cd), total salts (TS), total nitrogen (TN), total phosphorus (TP), total potassium (TK), soil organic carbon (SOC), and pH values. The analysis of soil samples was based on the industry standard “Analytical Methods for Regional Geochemical Samples” (DZ/T 0279-2016) [12] issued by the Ministry of Land and Resources of China; the analytical methods and detection limits are summarized in

Table 1. Test methods, detection limits, precision, recovery rates, and instrument models for the analysis of soil and plant samples.

Matrix	Parameter	Analytical Method	Detection Limit	Precision (%)	Recovery (%)	Instrument Model
Soil	As	XRF	0.2	3.2	98.5 ± 2.1	Shimadzu, Kyoto, Japan (EDX-8000, 2018)
	Hg	HG-AFS	0.005	5.8	102.3 ± 4.7	Jitian, Beijing, China (AFS-9700, 2019)
	Cr	ICP-MS	0.2	4.1	97.2 ± 3.0	Thermo Fisher, Waltham, MA, USA (iCAP RQ, 2020)
	Ni	ICP-MS	0.6	4.3	101.5 ± 3.5	Agilent, Santa Clara, CA, USA (7900 ICP-MS, 2019)
	Cu	ICP-MS	0.6	3.9	99.8 ± 2.8	PerkinElmer, Waltham, MA, USA (NexION 350D, 2018)
	Zn	ICP-MS	0.03	4.0	96.7 ± 3.2	PerkinElmer, Waltham, MA, USA (NexION 350D, 2018)
	Cd	GFAAS	0.05	4.5	103.4 ± 4.0	Analytik Jena, Jena, Germany (ZEEnit 700P, 2020)
	Pb	ICP-MS	0.5	3.7	98.1 ± 2.5	Thermo Fisher, Waltham, MA, USA (iCAP RQ, 2020)
	TS	Conductivity conversion method	20	2.5	100.8 ± 1.5	STEP Systems, Heidelberg, Germany (PNT3000, 2019)
	pH	Ion-selective electrode	0.01	0.8	—	Oakton, Vernon Hills, IL, USA (PCSTestr 35, 2018)
	TN	Kjeldahl digestion	0.01	2.2	99.3 ± 1.8	Elementar, Langenselbold, Germany (vario MACRO cube, 2019)
	TP	Molybdenum blue spectrophotometry	5–10	3.0	97.6 ± 2.4	Hach, Loveland, CO, USA (DR3900 Spectrophotometer, 2018)
	SOC	Potassium dichromate oxidation	0.1	2.8	101.2 ± 2.0	Elementar, Langenselbold, Germany (vario MACRO cube, 2019)
	TK	Flame photometry	0.2	1.9	98.9 ± 1.2	Sherwood, Cambridge, UK (Model 410 Flame Photometer, 2018)
Plant	As	HG-AFS	0.01	6.1	94.8 ± 5.2	Thermo Fisher, Waltham, MA, USA (iCE 3500, 2020)
	Hg	CVAAS	0.005	7.3	105.6 ± 6.8	Milestone, Milan, Italy (DMA-80, 2019)
	Cr	ICP-MS	0.01	5.2	97.5 ± 4.3	Agilent, Santa Clara, CA, USA (8900 ICP-MS + 1260 HPLC, 2020)
	Ni	GFAAS	0.05	6.5	102.4 ± 5.7	Thermo Fisher, Waltham, MA, USA (SOLAAR M6, 2019)
	Cu	ICP-OES	0.5	5.8	99.2 ± 4.9	PerkinElmer, Waltham, MA, USA (Avio 500, 2018)
	Zn	ICP-OES	0.1	4.7	96.3 ± 3.8	Agilent, Santa Clara, CA, USA (5110 VDV, 2019)
	Cd	GFAAS	0.01	6.9	101.7 ± 6.0	Thermo Fisher, Waltham, MA, USA (iCE 3500, 2020)
	Pb	ICP-MS	0.005	5.5	98.4 ± 4.5	Agilent, Santa Clara, CA, USA (7900 ICP-MS, 2020)

Notes: XRF: X-ray fluorescence spectrometry; HG-AFS: Hydride generation–atomic fluorescence spectrometry; ICP-MS: Inductively coupled plasma mass spectrometry; GFAAS: Graphite furnace atomic absorption spectrometry; CVAAS: Cold vapor atomic absorption spectrometry. As, Hg, Cr, Ni, Cu, Zn, Cd, Pb, and TS: mg/kg; TN, TP, TK, SOC: %; pH: Unitless.

. The precision and accuracy were primarily controlled using national first-grade standard reference materials, with recovery rates of 100% [13]. The reporting rates for individual elements in the soil samples ranged from 99.4% to 100%, with an overall element reporting rate of 100%, meeting the quality control requirements.

Additionally, for the typical regional vegetation, including P. euphratica, T. ramosissima, cotton, and jujube, root systems and rhizosphere soils were analyzed for eight heavy metals: As, Cr, Ni, Cu, Zn, Hg, Pb, and Cd, to further elucidate the migration and enrichment mechanisms of heavy metals in the soil–plant system. All analytical procedures were performed at the Analysis and Testing Center of Xinjiang Uyghur Autonomous Region Bureau of Nonferrous Geological Exploration. All analytical results passed quality acceptance, with each parameter meeting relevant regulatory standards. The determination of heavy metal concentrations in crops was carried out in accordance with the Food Safety National Standard—Determination of Total Mercury and Organic Mercury in Food (GB 5009.17-2021) [14] and Food Safety National Standard—Determination of Multi-elements in Food (GB 5009.268-2016) [15]; the analytical methods and detection limits are provided in Table 1. Analytical quality was controlled in accordance with the “Specification for Land Quality Geochemical Assessment” (DZ/T 0295-2016) [11]. For each batch of samples, 1–2 parallel standard reference materials of the same type were included and analyzed alongside the samples, and the relative error of a single determination for each sample was calculated; if the relative error was ≤30%, the sample was deemed acceptable. The precision for crop samples was controlled by duplicate analysis, with a duplication proportion of 100%; if the relative double difference for duplicate samples was ≤30%, it was considered acceptable. Calculations showed that the relative error and relative double difference for all elements were less than 30%, resulting in a 100% qualification rate, fully meeting quality control requirements. The analytical methods were selected in accordance with the national standard protocol DZ/T 0279-2016 [12] to ensure accuracy and comparability. Specifically, HG-AFS was employed for the determination of As and Hg due to its high sensitivity toward these elements; ICP-MS was used for Cd, Pb, Cr, and Ni owing to its low detection limits and ability to resolve potential interferences; and ICP-OES was applied for Cu and Zn as it provides robust and accurate performance for elements typically present at higher concentrations.

2.3. Analytical Methods

The Bioconcentration Factor (BCF) is a critical index for quantifying chemical accumulation in organisms [16]. It effectively evaluates a plant’s capacity to accumulate elements from the substrate and directly reflects its enrichment effect on specific heavy metals. The BCF is calculated as the ratio of the heavy metal concentration in the plant tissue (C_plant, mg/kg) to the total concentration of the corresponding heavy metal in the associated soil (C_soil, mg/kg). A higher BCF value indicates stronger enrichment capacity. The expression is:

BCF = C_plant/C_soil

(1)

Data analysis protocols: Descriptive statistics and Principal Component Analysis (PCA) were performed using SPSS 24 (IBM Corp., Armonk, NY, USA, released 2016). Bar charts, profile diagrams, and box plots were generated with Origin 2022 (OriginLab Corp., North Andover, MA, USA, released 2021). Gradient distribution maps of heavy metals were created using ArcGIS 10.2 (ESRI Inc., Redlands, CA, USA, released 2013). Redundancy Analysis (RDA) was conducted with Canoco 5.0 (Microcomputer Power, Ithaca, NY, USA, released 2012).

2.4. Geostatistical Analysis and Spatial Validation

To quantitatively characterize the spatial heterogeneity and autocorrelation of soil heavy metals and physicochemical properties, geostatistical analyses were performed using GS+ 9.0 (Gamma Design Software, Minneapolis, MN, USA, released 2008) and ArcGIS 10.2. The core analytical framework employed semivariogram modeling and ordinary kriging interpolation, validated through rigorous spatial cross-validation [17].

Semivariogram Modeling: The spatial dependence of each element was quantified by fitting experimental semivariograms to theoretical models [18]. The semivariance γ(h) at lag distance h was calculated as:

$$\mathsf{\gamma }\left(\mathrm{h}\right)={\displaystyle \frac{1}{2\mathit{N}\left(\mathit{h}\right)}}{\sum }_{\mathit{i}=1}^{\mathit{N}\left(\mathit{h}\right)}{\left[\mathit{Z}\left({\mathit{x}}_{\mathit{i}}\right)-\mathit{Z}\left({\mathit{x}}_{\mathit{i}}+\mathit{h}\right)\right]}^2$$

(2)

where N(h) is the number of sample pairs separated by distance h, and Z(x_i) is the measured value at location x_i. Three theoretical models (exponential, Gaussian, spherical) were evaluated for optimal fit for each individual element. The best-fitting model for a given element was selected based on achieving the minimized residual sum of squares (RSS) and the maximized coefficient of determination (R²) among the three candidate models. Key parameters derived from the selected model for each element included:

Nugget variance (C₀): Representing micro-scale variation or measurement error.

Sill (C₀ + C): Total variance at which the semivariogram plateaus.

Range (A₀): Distance beyond which spatial autocorrelation becomes negligible.

Nugget-to-sill ratio (C₀/(C₀ + C)): Classifying spatial dependence as strong (<25%), moderate (25–75%), or weak (>75%).

Kriging Interpolation and Validation: The optimal semivariogram model and its associated parameters (

Table 2. Geostatistical parameters and optimal semivariogram models for the spatial interpolation of soil heavy metals and total salts (TS).

Element	Theoretical model	Nugget (C0)	Sill (C0 + C)	Nugget Ratio (%)	Range (m)	R2	RSS	Confidence (%)
As	Index	0.02	0.14	14.28	4920	0.77	6.10 × 10−5	98%
Hg	Gauss	0.02	0.14	14.28	3222	0.21	1.08 × 10−4	84%
Cr	Index	0.01	0.06	16.67	4200	0.35	9.66 × 10−5	94%
Ni	Index	0.01	0.07	14.28	3570	0.22	4.01 × 10−5	85%
Cu	Index	0.01	0.12	8.33	2520	0.23	1.16 × 10−4	86%
Zn	Index	0.01	0.07	14.28	3240	0.24	4.20 × 10−5	95%
Cd	Gauss	0.01	0.06	16.67	2286	0.24	5.15 × 10−5	88%
Pb	Gauss	0.01	0.04	24.85	3186	0.38	8.35 × 10−6	92%
TS	Gauss	0.01	0.02	50	3984	0.21	1.76 × 10−8	82%

Note: C₀: representing micro-scale variation or measurement error; C₀ + C: representing the total variance; Nugget ratio (C₀/(C₀ + C)) classifies spatial dependence as strong (<25%), moderate (25–75%), or weak (>75%); Range (A₀) is the distance of spatial autocorrelation. R²: Coefficient of determination for the fitted model; RSS: Residual sum of squares; Confidence (%): Proportion of validation points within the 95% prediction interval during cross-validation.

) were input into ordinary kriging to generate continuous spatial distribution maps for all elements [19]. The interpolation estimator is defined as:

$$\widehat{\mathit{Z}}\left({\mathit{x}}_{\mathit{o}}\right)={\sum }_{\mathit{i}=1}^{\mathit{n}}{\mathsf{\lambda }}_{\mathit{i}}\mathit{Z}\left({\mathit{x}}_{\mathit{i}}\right)$$

(3)

where λ_i are weights derived from semivariogram models, constrained by ∑λ_i = 1. Spatial accuracy was validated Via leave-one-out cross-validation, with performance assessed by:

Determination coefficient (R²) between observed and predicted values.

Residual sum of squares (RSS): $RSS={\sum }_{\mathit{i}=1}^{\mathit{n}}[Z\left({\mathit{x}}_{\mathit{i}}\right)-\widehat{\mathit{Z}}\left({\mathit{x}}_{\mathit{o}}\right){]}^2$.

Confidence level (%): Proportion of validation points falling within 95% prediction intervals.

Table 2 summarizes the optimal semivariogram models and parameters for each element, validated through the above procedures. All analyses adhered to the quality control standards specified in Section 2.2. The relatively low R² values for most elements (except As) can be attributed to the complex spatial heterogeneity of the study area, influenced by both natural processes (e.g., alluvial deposition, wind erosion) and anthropogenic activities (e.g., localized irrigation, fertilization). Elements like Hg and Ni, with R² < 0.25, exhibit strong point-source pollution characteristics or high small-scale variability, which reduces the overall fit of the global semivariogram model. Despite the low R², the cross-validation confidence levels remained acceptable (>80% for all elements), indicating that the interpolation models still provided reliable spatial predictions for management purposes.

3. Results

3.1. Characteristics of Typical Elements in Surface Soils

Descriptive statistics of elements in the study area (

Table 3. Descriptive statistics of typical element concentrations and physicochemical properties in the surface soils (0–30 cm) of the study area (n = 323).

Element	Min	Max	Mean	SD	Skewness	Kurtosis	CV	Xinjiang Background Value
As	1.76	24.60	10.08	3.88	1.05	1.01	0.39	11.2
Hg	0.01	0.14	0.02	0.01	5.87	80.19	0.53	0.02
Cr	21.10	75.70	42.30	9.27	0.63	0.29	0.22	49.3
Ni	10.50	39.20	22.26	5.37	0.51	0.05	0.24	26.6
Cu	7.78	40.90	18.89	6.27	0.94	0.73	0.33	26.7
Zn	26.10	93.80	53.62	13.10	0.45	-0.11	0.24	68.8
Cd	0.06	0.66	0.14	0.04	3.32	31.06	0.31	0.12
Pb	10.08	32.58	17.62	2.83	0.98	1.83	0.16	19.4
TS	0.26	393.00	21.32	36.33	4.18	26.98	1.70	—
pH	7.47	9.84	8.59	0.30	0.15	0.56	0.03	—
TN	0.01	0.16	0.04	0.02	1.15	2.45	0.48	—
TP	281.00	1213.00	572.58	119.20	1.22	3.43	0.21	—
SOC	0.05	2.17	0.47	0.30	1.24	2.56	0.64	—
TK	1.23	3.10	2.24	0.24	0.03	1.50	0.11	—

Note: Min: Minimum value; Max: Maximum value; Mean: Arithmetic mean calculated as the sum of all values divided by the number of samples (n = 323); Standard Deviation (SD): A measure of data dispersion around the mean, calculated as the square root of the variance, representing the dispersion of data around the mean; Skewness: A measure of the asymmetry of the probability distribution; Kurtosis: A measure of the “tailedness” of the probability distribution; CV: Coefficient of Variation (SD/Mean), expressed here as a decimal to quantify relative spatial variability. Xinjiang Background Value is provided for reference [21,22]. TS: Total Salts.

) reveal significant variations in their concentrations. The mean concentrations of As, Cr, Ni, Cu, Zn, and Pb were 10.08, 42.30, 22.26, 18.89, 53.62, and 17.62 mg/kg, when evaluated against China’s mandatory Soil Environmental Quality Risk Control Standard for Agricultural Land (GB 15618-2018) [20], all elements remain substantially below the risk screening thresholds for agricultural soils, confirming minimal contamination risk under current concentrations, respectively, all below the Xinjiang background values (11.2, 49.3, 26.6, 26.7, 68.8, and 19.4 mg/kg), suggesting these elements are primarily governed by natural pedogenic parent materials [21]. In contrast, Hg and Cd exhibited mean concentrations of 0.02 mg/kg and 0.14 mg/kg, slightly exceeding their background values (0.02 and 0.12 mg/kg), indicating potential anthropogenic inputs [22]. The coefficient of variation (CV) further elucidates spatial patterns: Cr, Ni, Zn, and Pb displayed low spatial variability (CV = 0.16–0.24), reflecting homogeneous distribution typical of natural sources, while As, Cu, and Cd showed moderate variability (CV = 0.31–0.39), likely attributable to localized anthropogenic activities superimposed on natural processes [23]. Notably, Hg exhibited high variability (CV = 0.53) with extreme skewness (5.87) and kurtosis (80.19), signaling potential point-source pollution. The mean TS concentration was 21.32 g/kg, classified as severe salinization according to Xinjiang criteria, consistent with prior findings [24], with its high CV (1.70) underscoring pronounced spatial heterogeneity.

Cross-validation confirmed model reliability (Table 2). As prediction achieved high accuracy (R² = 0.77, RMSE = 1.2 mg/kg), while prediction accuracy varied for other elements (e.g., Hg R² = 0.21), consistent with their spatial complexity. Despite lower R² for Hg and Ni attributed to complex source pathways, spatial distribution confidence levels exceeded 80% for all elements (e.g., As: 98%, Zn: 95%), confirming robust interpolation accuracy, with Hg notably meeting acceptance thresholds (RMSE = 0.008 mg/kg < detection limit). Nugget ratios for heavy metals (C₀/(C₀ + C)) were consistently below 25% (e.g., As: 14.28%, Pb: 24.85%), indicating strong spatial autocorrelation dominated by natural processes, while pH exhibited moderate correlation (50.00%).

Kriging interpolation revealed distinct spatial patterns (Figure 2). Naturally dominated elements (Cr, Ni, Zn, Pb) showed relatively homogeneous distributions [25], aligned with Quaternary sediments (e.g., localized Cr hotspots >60 mg/kg potentially linked to Tianshan-derived aeolian minerals) [26]. Anthropogenically influenced elements exhibited specific enrichment: As showed punctate enrichment near downstream irrigation and settlements (>0.05 mg/kg), Cd exceeded 0.5 mg/kg in long-cultivated irrigated areas (hotspots congruent with crops) [27], and Hg enrichment spatially correlated with As hotspots, suggesting shared sources like coal combustion. Zn distribution showed transitional patterns [28].

3.2. Soil–Vegetation Heavy Metal Enrichment Characteristics

Analysis of heavy metal concentrations in the rhizosphere soils of four typical plant species—P. euphratica, T. ramosissima, cotton, and jujube—revealed significant vegetation-type-dependent accumulation patterns in Figure 3. Quantitative measurements demonstrated that within natural vegetation systems, P. euphratica rhizosphere soils contained consistently lower concentrations of As (0.072–0.56 mg/kg), Cr (0.31–1.19 mg/kg), and Pb (0.17–0.66 mg/kg) compared to T. ramosissima, which exhibited elevated levels of 0.36–2.12 mg/kg for As, 0.66–6.11 mg/kg for Cr, and 0.32–3.00 mg/kg for Pb. Notably, T. ramosissima displayed markedly stronger accumulation capacity for Cd at 0.0125–0.5490 mg/kg and Hg at 0.010–0.078 mg/kg than P. euphratica. The high CV exceeding 1.0 suggest T. ramosissima possesses heightened sensitivity to localized pollution sources, as documented by Qian [29]. Critical anomalies were observed at specific sampling sites where Cr concentrations reached 6.11 mg/kg and Pb attained 3.00 mg/kg, potentially attributable to T. ramosissima’s deep-rooting physiology and saline–alkaline adaptation mechanisms described by Xue [30]. In contrast, economically cultivated vegetation exhibited distinct rhizospheric distribution patterns. Jujube demonstrated significantly higher enrichment of Cu at 4.12–12.0 mg/kg and Zn at 8.74–75.7 mg/kg compared to cotton, which showed respective ranges of 3.21–6.23 mg/kg for Cu and 8.79–19.5 mg/kg for Zn. This divergence likely stems from prolonged application of Cu and Zn-containing fertilizers in jujube cultivation zones. Meanwhile, cotton rhizosphere soils contained elevated Hg concentrations of 0.018–0.488 mg/kg, reflecting residual impacts from historical pesticide applications as reported by Bokossa [31].

Comparative analysis reveals that natural vegetation (P. euphratica, T. ramosissima) and economic crops (Cotton, Jujube) display distinct heavy metal content profiles in their rhizosphere soils. Natural vegetation exhibits higher coefficients of variation in heavy metal concentrations, indicating a greater sensitivity to environmental changes, whereas the distribution patterns of heavy metals in cultivated crops are more strongly influenced by anthropogenic management practices. Although the Cd levels in certain Jujube sampling points (0.0064–0.0442 mg/kg) are below safety thresholds, the maximum values approach regulatory limits for food safety, warranting focused attention.

3.3. Differential Heavy Metal Enrichment Patterns via BCF

Statistical analyses Via one-way ANOVA with Tukey’s post hoc test at α = 0.05 significance level quantified divergent heavy metal enrichment strategies as shown in Figure 4, where distinct superscript letters denote statistically homogeneous subgroups. P. euphratica demonstrated Cd hyperaccumulation (mean BCF = 1.59 ± 0.21, subgroup a), significantly surpassing (p < 0.05) T. ramosissima (0.80 ± 0.15, b), cotton (0.21 ± 0.05, bc), and jujube (0.12 ± 0.03, c) [32]. For Zn, P. euphratica maintained dominance (BCF = 2.41 ± 0.45, a), significantly exceeding (p < 0.05) all others (b: T. ramosissima 0.42 ± 0.08, cotton 0.25 ± 0.06, jujube 0.42 ± 0.09) [33].

Hg enrichment showed no significant inter-group differentiation (p > 0.05), all species sharing subgroup a (cotton highest: 2.15 ± 0.40). T. ramosissima exhibited significantly higher (p < 0.05) As accumulation (BCF = 0.06 ± 0.02, a) than P. euphratica (0.03 ± 0.01, b); cotton (0.05 ± 0.02, ab) was intermediate [34].

Pb accumulation formed a gradient: T. ramosissima (a: 0.05 ± 0.01) > cotton (bc: 0.04 ± 0.01) > P. euphratica (c: 0.02 ± 0.01) (p < 0.05); jujube (ab: 0.03 ± 0.01) bridged cotton and P. euphratica [35]. Ni: T. ramosissima (a: 0.04 ± 0.01) > P. euphratica (bc: 0.02 ± 0.01) and jujube (c: 0.02 ± 0.01) (p < 0.05). Cr: T. ramosissima achieved exclusive subgroup a status (0.02 ± 0.01), significantly higher (p < 0.05) than others. Cu accumulation remained uniformly low (<0.10 BCF) across all vegetation (a) [36].

These differential strategies establish P. euphratica as optimal for Cd and Zn remediation through its elevated BCF values in exclusive a subgroups. T. ramosissima broad-spectrum accumulation of As at 0.06, Pb at 0.05, and Ni at 0.04 in a subgroup levels warrants caution in multi-metal zones, while jujube’s Cd accumulation in c subgroup at 0.12 ± 0.03 demands cultivation restrictions near the 0.05 mg/kg food safety threshold [37].

3.4. Driving Factors

RDA combined with significance testing Via Monte Carlo permutation tests (999 permutations, p < 0.01 threshold) revealed that soil physicochemical properties significantly governed heavy metal migration and enrichment processes, with a cumulative explanatory power of 94.42%, highlighting their pivotal role in shaping spatial distribution patterns. This permutation procedure assessed the statistical significance of both the overall RDA model and the individual canonical axes by randomly reordering the sample-environmental variable relationships 999 times to generate a null distribution; the proportion of permutations yielding a test statistic (trace statistic) equal to or greater than the observed value determined the p-value, confirming the relationships were not due to chance [38]. Environmental factors drove spatial differentiation through orthogonal biaxial drivers [39]. RDA Axis 1 (variance contribution: 57.59%) primarily represented suppressive effects. pH (loading = −0.72, p < 0.01) and TS (loading = −0.65, p < 0.01) showed significant negative correlations (p < 0.01) with As, Cd, Cu, and Zn. In alkaline high-salinity zones (pH = 7.47–9.84, TS ≤ 393 g/kg; 37.1% of samples), heavy metal concentrations decreased by 42–68% compared to background values, aligning spatially with low-accumulation areas (Section 3.1). Chemically, alkaline conditions promote hydroxide precipitation of metal ions, while high salinity elevates ionic strength, suppressing bioavailability Via ion exchange processes [40].

RDA Axis 2 (variance contribution: 36.83%) reflected promotive effects. TN (loading = 0.51, p < 0.05), TP (loading = 0.63, p < 0.05), and SOC (loading = 0.58, p < 0.05) positively correlated with Cr, Ni, and Pb enrichment. Notably, significant synergy existed between TP and Cd (Pearson’s r = 0.42, p = 0.008), likely due to phosphorus compounds altering Cd speciation, enhancing its solubility and mobility (Figure 5).

Independent explanatory power of environmental factors ranked as follows: TK (29.3%) > SOC (19.8%) > TP (15.3%) > TS (14.9%) > pH (14.2%) > TN (7.1%) (

Table 4. Results of redundancy analysis (RDA) forward selection showing the importance of environmental factors in explaining the variance of heavy metal distribution.

Feature	Explained Variance (%)	Contribution (%)	F-Value	p-Value
TK	29.3	37.4	14.7	0.002
SOC	19.8	21.6	10.2	0.008
TP	15.3	13.3	6.4	0.016
TS	14.9	11.2	7.3	0.022
pH	14.2	8.9	6.0	0.038
TN	7.1	7.6	3.4	0.048

Note: Explained Variance (%): The proportion of variance in the heavy metal data explained by each environmental factor alone. Contribution (%): The conditional (unique) contribution of each factor to the explained variance of the constrained axes. F-value: The F-statistic from the Monte Carlo permutation test. p-value: The significance level of the factor’s contribution (p < 0.01 is highly significant). TK: Total Potassium; SOC: Soil Organic Carbon; TP: Total Phosphorus; TS: Total Salts; TN: Total Nitrogen.

). TK demonstrated the highest conditional effect (37.4%) on constrained axes and exhibited significant positive correlations with Cd and Zn enrichment (p = 0.002), indicating potassium-enhanced accumulation of these metals. Mechanistically, K⁺ ions act as strongly exchangeable cations that displace adsorbed Cd²⁺/Zn²⁺ from soil colloids through cation exchange, increasing their mobility and bioavailability [41]. Concurrently, KCl-based fertilizers may introduce Cd/Zn contaminants, driving co-variation with TK [42]. SOC concentrations (0.05–2.17 g/kg) positively regulated Cr and Pb migration (p = 0.008) Via formation of soluble complexes between oxygen-containing functional groups (e.g., carboxyl, phenolic hydroxyl) and metal ions [43]. Notably, synergistic suppression by TS and pH intensified in extreme zones (TS > 350 g/kg, pH > 8.5), where hypersaline–alkaline conditions systemically reduced metal activity through precipitation as carbonates/hydroxides and enhanced adsorption onto soil particles [44], validating RDA Axis 1′s suppression effects.

4. Discussion

4.1. Synergistic Salinity–Alkalinity Suppression Mechanism on Heavy Metals

This study establishes a spatially and chemically validated closed-loop evidence chain for the salinity–alkalinity co-suppression mechanism. Spatial validation Via high-confidence interpolation (mean = 90% for geogenic elements; Table 2) confirmed homogeneous Cr, Ni, and Zn distributions consistent with Quaternary sediment weathering, while Hg’s low R² (0.21) reflected point-source anthropogenic enrichment—highlighting the need for targeted sampling in hotspots (e.g., settlements, irrigated zones). Crucially, in alkaline high-salinity zones (pH = 7.47–9.84, TS ≤ 393 g/kg; 37.1% of samples), As, Cd, Cu, and Zn decreased by 42–68% versus Xinjiang baselines [45]. These suppression areas align spatially with alluvial fan marginal depressions, demonstrating hydrogeological control over salinity–alkalinity synergy. RDA further confirmed dual-driven suppression: Axis 1 revealed significant negative correlations (p < 0.01) between heavy metals and pH (loading = −0.72)/TS (loading = −0.65).

Chemically, suppression operates through: (i) Precipitation-immobilization at pH > 8.5, forming low-solubility phases (e.g., Cu(OH)₂, CdCO₃) that reduce bioavailability; (ii) Leaching-migration, where high salinity (TS > 200 g/kg) enhances competitive displacement of Cd²⁺/Zn²⁺ by Na⁺/Ca²⁺ from adsorption sites, coupled with downward solute transport via irrigation infiltration [2].

While these macro-scale processes are evident, micro-scale analyses from analogous environments provide direct evidence for the immobilization mechanisms. Isotopic tracing studies (e.g., using ¹¹⁰Cd or ⁶⁷Zn) have demonstrated that a significant fraction of added Cd and Zn in saline–alkaline soils is rapidly transferred into less labile pools, with reduced uptake by plants, supporting the precipitation and adsorption pathways inferred here [46]. Furthermore, molecular-scale investigations using synchrotron-based techniques have identified the specific speciation changes: under high pH, Zn is often found incorporated into hydrozincite [Zn₅(CO₃)₂(OH)₆], while Cd forms otavite (CdCO₃) or is sequestered into the structure of Fe/Mn (oxyhydr)oxides that are stable under alkaline conditions. The high ionic strength from salts compresses the electrical double layer, reducing the electrostatic potential and enhancing the inner-sphere complexation of metal ions onto mineral surfaces, a process validated by surface complexation modeling and spectroscopic evidence [47].

While analogous mechanisms occur in saline systems, their dominance in this arid ecosystem is evidenced by system-wide suppression—reflected in homogeneous geogenic distributions (Cr/Ni/Zn) and reinforced by spatial confidence metrics and is consistent with the micro-scale processes identified in other studies.

4.2. Vegetation-Type-Driven Differentiation in Heavy Metal Enrichment

This study reveals fundamentally distinct heavy metal accumulation patterns between natural vegetation and economic crops, governed by synergistic interactions of physiological adaptations and anthropogenic interventions. P. euphratica likely employs a metal exclusion strategy via its deep root systems (reaching 5–8 m depth), possibly accessing less contaminated water sources or avoiding surface pollution hotspots, which may contribute to its significantly lower rhizospheric As, Cr, and Pb concentrations compared to T. ramosissima. This deep-rooting avoidance mechanism presents a contrasting strategy to the surface root accumulation commonly reported for herbaceous hyperaccumulators in humid regions [48], highlighting how phytoremediation strategies must be ecosystem-specific. In contrast, T. ramosissima mobilizes As, Cr, and Pb via saline–alkaline adaptation mechanisms, wherein root-exuded organic acids elevate rhizosphere pH to 8.2–9.0. This process explains critical concentration anomalies reaching 6.11 mg/kg for Cr and 3.00 mg/kg for Pb, with coefficients of variation exceeding 1.0 confirming its utility as a pollution sentinel species. Its role as an accumulator in arid saline soils is notable when compared to other common riparian species globally, which may not exhibit the same tolerance or accumulation capacity under such extreme stress [49].

BCF quantification demonstrates P. euphratica’s Cd and Zn hyperaccumulation capability, evidenced by BCF values of 1.59 and 2.41, respectively, attributable to vacuolar compartmentalization through upregulated HMA₃ transporter expression [50]. While hyperaccumulation of Cd and Zn is well-documented in species like Sedum alfredii and Noccaea caerulescens, their application is often limited to acidic or neutral, non-saline soils [48]. The discovery of strong accumulation traits in P. euphratica is significant because it offers a native, phreatophytic tree species capable of remediating deeper soil profiles in arid, saline environments where conventional hyperaccumulators fail to thrive. T. ramosissima exhibits broad-spectrum accumulation of As, Pb, and Ni, necessitating caution in multi-metal contaminated zones. This multi-metal accumulation pattern differs from the more element-specific accumulation observed in many studied halophytes, underscoring the unique ecotoxicological risk and potential utility of T. ramosissima [51]. Among economic crops, jujube shows pronounced Cu and Zn enrichment in rhizosphere soils, with maxima of 12.0 mg/kg and 75.7 mg/kg, respectively, directly linked to prolonged foliar fertilization. However, Cd concentrations approaching the 0.05 mg/kg food safety threshold in jujube roots stem from phosphorus fertilizer-induced Cd mobilization, a widespread phenomenon also observed in orchards worldwide [52]. Cotton accumulates Hg predominantly from historical pesticide applications, yielding the highest BCF of 2.15 due to shallow root confinement in surface soils [53]. This high transfer factor for Hg in cotton is particularly alarming when compared to other fiber crops, suggesting a crop-specific vulnerability that necessitates targeted management [54].

Consequently, a three-tiered management framework emerges: P. euphratica serves optimally for Cd and zinc phytoextraction in arid regions, providing a novel alternative to traditional hyperaccumulators in this challenging environment, while T. ramosissima supports As and Pb phytostabilization but requires concurrent pollution monitoring. Jujube cultivation zones demand immediate biochar soil amendment for Cd immobilization and agricultural land-use restrictions, a recommendation supported by successful trials in other tree crop systems [55]. Cotton fields necessitate pesticide phase-outs complemented by microbial Hg detoxification technologies to achieve integrated contamination control and ecological security [56].

4.3. Multidimensional Control Framework for Saline Farmland Heavy Metal Risks

Integrating spatial heterogeneity, plant enrichment, and driving mechanisms, we propose a ‘soil–vegetation–environment’ co-management framework. This integrated approach addresses a gap in remediation strategies, which often focus solely on either soil amendments or plant selection, particularly in arid saline environments where such holistic frameworks are scarce [57]. For soil chemistry, high-pH (>8.5) and high-TS (>200 g/kg) zones naturally suppress metal activity, while moderate-/low-salinity areas (pH < 8.0, TS < 50 g/kg) require mitigation of phosphorus–cadmium synergy using Fe-based immobilizers. The proposed dual-threshold (salinity/pH and TP) system offers a more nuanced management tool compared to single-factor guidelines often used in non-saline agricultural lands [58]. In vegetation allocation, Cd/Zn-contaminated zones should prioritize planting Cd- and Zn-tolerant species (e.g., long-cultivated irrigated areas) to remediate deep pollution, whereas cotton should be avoided in high-Hg soils and jujube cultivation in high-TP soils warrants caution due to observed rhizospheric Cd accumulation. This species-specific zoning recommendation, based on quantitative BCF comparisons, moves beyond generic advice and provides actionable guidance tailored to the major crops and native vegetation of the Tarim Basin, a refinement over more general phytomanagement proposals [59].

For environmental early-warning, we recommend monitoring salinity–alkalinity suppression thresholds and phosphorus-cadmium risk thresholds to trigger management actions. This framework, built on driver weights from the RDA and spatial validation results, is conceptually feasible; however, its practical implementation must explicitly address socio-economic trade-offs. Key considerations include: (i) crop substitution and rotation decisions that affect farm income, market access, and risk profiles; (ii) costs and benefits of remediation measures (capital and recurrent costs, labor, equipment, potential yields changes, and environmental externalities) and their payback periods; (iii) access to finance, extension services, and policy incentives that influence adoption and continuity; (iv) equity and distributional effects among smallholders versus larger producers; (v) ongoing monitoring costs and data requirements; and (vi) policy instruments (subsidies, incentives, penalties) and their fiscal sustainability. To ensure economic viability alongside environmental efficacy, we recommend integrating an economic evaluation (e.g., cost-benefit analysis, MCDA, or LCCA) with the framework to identify context-specific, cost-effective, and socially acceptable management options. Practical implementation should proceed Via pilot demonstrations, stakeholder engagement, and data collection on crop yields, market prices, input and remediation costs, and program administration costs, enabling iterative refinement of both the ecological model and the socio-economic assessment.

4.4. Limitations and Future Research Directions

While this study elucidates the key mechanisms and patterns of heavy metal behavior in the arid saline soils of the Tarim Basin, several limitations should be acknowledged. Firstly, the findings are based on a single sampling campaign, which provides a robust spatial snapshot but cannot capture the temporal dynamics of metal mobility and plant uptake in response to seasonal variations in groundwater level, irrigation practices, and evaporation intensity. Future research should implement long-term monitoring across different seasons to quantify these temporal fluxes and validate the stability of the proposed immobilization thresholds. Secondly, although the plant sample size (n = 55) is sufficient for identifying significant interspecies differences, a larger sample set, particularly for economic crops, would enhance the statistical power for intraspecies variability analysis and more robust risk assessment models.

A critical limitation is the absence of direct measurement of heavy metal concentrations in the edible parts of crops (e.g., jujube fruits, cotton seeds). Our discussion on food safety risk for jujube is based on rhizosphere soil and root data, while the actual human exposure pathway is through the consumption of edible tissues. Therefore, future work must prioritize analyzing metal translocation factors (from root to shoot to fruit) to accurately assess the potential health risks and establish reliable safety thresholds for agricultural products. Finally, while we have inferred molecular-scale immobilization mechanisms (e.g., precipitation, adsorption) from our macro-scale data and literature, direct evidence is lacking. Subsequent investigations should employ advanced micro-spectroscopic techniques (e.g., synchrotron-based X-ray absorption spectroscopy-XAS) to precisely characterize the solid-phase speciation of metals (e.g., Cd, Zn) in saline–alkaline soils and isotopic tracing (e.g., ¹¹⁴Cd) to quantify the fluxes and pathways of metal uptake in the root systems of key species like P. euphratica and T. ramosissima. Addressing these limitations will be crucial for transforming the proposed management framework from a conceptual model into a predictive and optimized tool for ecological security and agricultural safety in arid regions.

5. Conclusions

This study systematically investigates the behavior mechanisms and plant enrichment characteristics of heavy metals in the arid saline–alkaline soils of the Tarim Basin. The main conclusions are as follows:

(1)

Soil salinity and alkalinity significantly inhibit the mobility of heavy metals. When TS > 200 g/kg and pH > 8.5, the bioavailability of metals such as As, Cd, Cu, and Zn decreases by 42–68%. The mechanisms primarily include the formation of low-solubility precipitates and competitive displacement of heavy metal ions from adsorption sites by Na⁺/Ca²⁺.

(2)

Plant type significantly influences heavy metal enrichment strategies. P. euphratica exhibits hyperaccumulation capacity for Cd and Zn (BCF > 1), demonstrating potential for targeted phytoremediation, while T. ramosissima tends to accumulate As and Pb, requiring enhanced ecological risk monitoring in multi-metal polluted areas.

(3)

Economic crops show notable heavy metal transfer risks. Cotton has the highest enrichment factor for Hg (BCF = 2.15), and jujube roots approach Cd safety thresholds in phosphorus-rich soils, indicating the need for crop-specific pollution control measures.

(4)

Key factors driving heavy metal behavior in soil include: pH and TS jointly inhibiting metal activity, while TP and SOC enhance availability through complexation and phosphorus-cadmium synergy.

Based on these findings, a “soil–vegetation” co-management dual-threshold framework is proposed: utilizing natural immobilization capacity in high saline–alkaline areas, implementing phosphorus control and passivation technologies in moderate to low saline–alkaline areas, and spatially distinguishing priority zones for phytoremediation and agricultural risk management. This provides a theoretical basis and practical pathway for the sustainable management of degraded farmland and ecosystems in the region.