Integrated Geochemical, Vegetation, and Risk Assessment of a Pb–Zn Slag Reprocessing Site in Southern Kazakhstan: Implications for Sustainable Remediation Prioritization

Abstract

Reprocessing historical lead–zinc (Pb–Zn) slag offers a circular-economy pathway for secondary metal recovery, yet it can remobilize legacy contaminants where containment is inadequate, transferring risk to the surrounding land. Sustainable management of such sites requires frameworks that link contamination assessment to actionable remediation. We integrated ICP-OES geochemistry, native-plant biomonitoring, and US EPA RAGS-based risk modeling at an active Pb–Zn slag reprocessing site in Shymkent, Southern Kazakhstan. Twenty-four soil samples along four cardinal transects, two reference samples, and four composite plant samples (Centaurea pseudosquarrosa + Plantago lanceolata) were analyzed for ten metals by ICP-OES. UCC-referenced indices classified six metals as geoaccumulation Class 6 at most points (enrichment factors up to 90,871, confirming an exclusively anthropogenic origin). Peak concentrations reached 9350 mg·kg⁻¹ Pb, 290 mg·kg⁻¹ Cd, and 10,900 mg·kg⁻¹ As—exceeding Kazakhstan MPC by 72×, 290×, and 5450×. Worst-case carcinogenic risk reached 4.3 × 10⁻³ (43× above the US EPA threshold), driven almost entirely by arsenic (93%); ecosystem risk (RCR_total = 223) was dominated by cadmium (43%), arsenic (27%), and mercury (16%)—a disconnect between mass-based and toxicity-based prioritization. On this basis we propose a three-tier remediation framework (engineered containment, phytostabilization, monitored attenuation) that couples resource recovery with contamination control, is transferable to analogous Pb–Zn legacy sites, and supports sustainable land use, urban resilience, and responsible secondary-resource use.

1. Introduction

Reprocessing historical lead–zinc (Pb–Zn) slag remains a globally significant environmental challenge. Slag heaps and tailings impoundments, accumulated over decades of pyrometallurgical operations, act as persistent metal reservoirs, releasing Pb, Zn, Cd, As, and Hg into surrounding soils via wind-driven particle transport, surface runoff, and slow leaching into groundwater [1,2,3]. Although secondary recovery of residual metals from such slags is increasingly recognized as a circular-economy opportunity [4,5], reprocessing operations can remobilize legacy contaminants if dust suppression, runoff control, and surface stabilization are inadequate [6,7]. This duality—slag as both a valuable secondary raw material and a chronic source of soil-borne metal release—underscores the need for site-specific environmental assessment as an integral component of any reprocessing program. Where engineered containment, dust control, and runoff management are absent or insufficient, the same operations that close the metal loop can paradoxically extend the contamination footprint of the original smelting legacy, transferring risk from the heap itself to surrounding agricultural and residential land.

Southern Kazakhstan, specifically the Turkestan Region, hosts the Achisai polymetallic ore field and the Shymkent lead–zinc processing complex, together forming one of the largest concentrations of Pb–Zn metallurgical legacy in Central Asia [8]. The former Shymkent lead plant operated for about seven decades before Yuzhpolymetal JSC closed in 2012, leaving an uncontained slag heap that remains subject to aeolian dispersion under the semi-arid climate [6,8]. For regional context, the neighboring single-industry town of Kentau, 60 km to the south, which shares the same Pb–Zn ore–processing legacy, still had elevated soil Pb and Cd levels after more than two decades of passive recovery [9], underscoring the persistence of smelter contamination across the Turkestan Region. Recent assessments across Kazakhstan have likewise documented elevated soil heavy-metal burdens at industrial and urban sites [10,11], and Monte Carlo-based health risk modeling in comparable Pb–Zn mining areas confirms cadmium and arsenic as dominant carcinogenic risk drivers despite the conventional emphasis on lead [12].

To contextualize analytical data from such legacy sites, six complementary normalized indices are conventionally used. The contamination factor (CF) and the geoaccumulation index (Igeo) express contamination relative to a geochemical background; the enrichment factor (EF) distinguishes anthropogenic enrichment from lithogenic input; the pollution load index (PLI) and the modified degree of contamination (mCd) summarize multi-element loading; and the potential ecological risk index (RI) translates these into an integrated ecosystem-hazard score [13]. In situ phytoindication, based on tissue concentrations in native vegetation, provides independent biological validation of geochemical findings and identifies candidate species for phytostabilization [14].

Existing assessments of the Achisai–Kentau metallurgical corridor have examined individual aspects—soil chemistry, atmospheric deposition, or human exposure—in isolation. The present study integrates ICP-OES geochemistry along four cardinal-direction transects, a six-index pollution framework referenced to the upper continental crust (UCC), composite plant-tissue biomonitoring of two native species (Centaurea pseudosquarrosa and Plantago lanceolata), and US EPA RAGS-based health and ecosystem risk modeling into a single, site-specific analysis. To our knowledge, this combination has not been applied previously at the Shymkent slag reprocessing site, and the resulting framework is directly transferable to analogous Pb–Zn legacy reprocessing sites across Central Asia and globally. By translating contamination evidence into a tiered remediation strategy that reconciles secondary-resource recovery with contamination control, the study addresses the sustainability dimension of slag reprocessing and contributes directly to the United Nations Sustainable Development Goals on good health and well-being (SDG 3), sustainable cities and communities (SDG 11), responsible consumption and production (SDG 12), and life on land (SDG 15).

The specific objectives of this study were therefore to: (i) quantify Pb, Zn, Cd, Cu, As, Hg, Cr, Ni, Mn, and Fe concentrations in soils along four cardinal-direction transects (0–100 m) and at two distal reference points (300 m and 1.5 km) at an active Pb–Zn slag reprocessing site in Shymkent, Southern Kazakhstan; (ii) classify the resulting contamination using six complementary geochemical indices (CF, Igeo, EF, PLI, mCd, RI) referenced to the upper continental crust; (iii) characterize community-level metal uptake in two native species (Centaurea pseudosquarrosa and Plantago lanceolata) via the bioconcentration factor; (iv) quantify human-health and ecosystem risks using US EPA RAGS and REACH PNEC-based RCR frameworks; and (v) translate the integrated evidence into a site-specific, three-tier remediation prioritization framework transferable to analogous Pb–Zn legacy reprocessing sites in Central Asia and globally.

Comparable site-resolved assessments of Pb–Zn smelter and slag-reprocessing legacy contamination have been reported across multiple continents. Yang et al. [15] characterized the distribution of heavy metals in topsoil around a Pb–Zn smelter in Henan Province (Central China), reporting Pb, Cd, and As enrichment levels that consistently exceeded national risk-screening thresholds. Ettler et al. [16] used Pb isotopic evidence to trace the dispersal of contamination in stream sediments from the Příbram mining and smelting district (Czech Republic), demonstrating that smelter-derived Pb persists in the fluvial environment decades after operational closure. Angelovičová et al. [17] documented suppression of soil enzymatic activity in the Middle Spiš mining area (Slovakia), and Tembo et al. [18] quantified Cu–Pb–Cd–Zn contamination around the Kabwe smelter (Zambia)—a site that consistently ranks among the world’s most polluted urban environments. More recently, Wang et al. [19] applied Monte Carlo simulation to characterize health risks in the Xicheng Pb–Zn mining area (China), identifying Cd and as dominant carcinogenic drivers despite a smaller absolute Pb burden. In post-Soviet Central Asia, Junusbekov et al. [9] documented elevated heavy-metal levels in soils around the Kentau Pb–Zn complex (~150 km north of Shymkent), Tleuova et al. [20] reported hydrogeological controls on heavy-metal mobility in Southern Kazakhstan, and Faurat et al. [10] characterized snow–soil–vegetation transfer of heavy metals in an industrial city in Kazakhstan. Across these geographically diverse cases, three recurring patterns emerge: (i) the spatial footprint of contamination consistently extends well beyond the immediate slag-deposit perimeter; (ii) Cd and As, rather than the more abundant Pb and Zn, often dominate human-health and ecosystem risk; and (iii) integrated assessments combining geochemistry, plant biomonitoring, and quantitative risk modeling remain rare relative to single-component studies. The present study addresses this last gap for the Achisai–Shymkent corridor of Central Asia, which to date has been examined only through fragmented single-aspect assessments [11].

2. Materials and Methods

2.1. Study Area

The study was conducted at and around an active lead–zinc slag reprocessing facility in Shymkent, Southern Kazakhstan (42.30° N, 69.58° E; Figure 1). The site lies within the semi-arid Turkestan Region (Köppen BSk), which has a strongly continental climate with a mean annual precipitation of about 400 mm and prevailing easterly and westerly winds that control the directional dispersion of contaminated dust from the slag heap [6,8].

The slag heap occupies a substantial fraction of the former Shymkent lead-smelting site and consists primarily of fine-grained (0.5–15 mm), dark gray to black, granular particles. Non-ferrous smelting slags of this type are characterized by complex mineralogy (pyroxenes, olivine, melilite, magnetite, residual galena, and sphalerite) and by the capacity to release Pb, Zn, and Cd upon weathering, with release rates sensitive to particle size and pH [21]. SEM-EDS analysis (JEOL JSM-6490LV, Tokyo, Japan) confirmed the presence of PbS, ZnS, ZnO, PbO, and Fe₂O₃ phases. Slag pH ranged from 7.5 to 8.5. The Badam River flows approximately 20 m west of the facility boundary, providing a potential pathway for metal transport into the aquatic environment [20]. In the reach adjacent to the slag-reprocessing site, the Badam River is a shallow mid-order stream characteristic of the semi-arid Turkestan piedmont, with seasonally variable hydrology typical of foothill rivers in Southern Kazakhstan [20]. Water depth and channel width vary substantially between the dry season and the spring snowmelt freshet (April–May), with deeper and wider flow conditions during the freshet that increase the river’s capacity to mobilize and transport fine particulate material delivered from adjacent upland surfaces. Hydrogeologically, the site overlies a shallow Quaternary alluvial aquifer hosted in poorly sorted gravels and sands of relatively high effective porosity, which provides limited natural attenuation capacity for dissolved metal species [20]. Two principal pathways for heavy-metal migration from the slag heap to the aquatic environment are therefore operative: (i) episodic high-intensity precipitation events generate surface runoff that can transport fine slag particulates and metal-bearing leachate across the unprotected slag-heap perimeter toward the riverbank within hours; and (ii) slow vertical infiltration through the vadose zone may deliver dissolved Cd, Zn, and other mobile species to the underlying shallow aquifer, with subsequent baseflow discharge to the river over longer timescales. These hydrogeological conditions, combined with the absence of an engineered containment liner beneath the slag heap, define the principal exposure pathways for off-site metal migration and provide the geomorphological rationale for the riverbank-directed (westward) component of the transect sampling design used in this study.

2.2. Soil and Vegetation Sampling

Soil and plant tissue samples were collected on 24 April 2026 in accordance with GOST R 58595-2019 (soil sampling) [22], GOST 17.4.3.01-2017 [23] (general sampling requirements), and GOST 17.4.4.02-2017 (sample preparation) [24]. Twenty-four soil samples (0–50 cm depth) were collected along four cardinal-direction transects (north, south, east, west) at six distances from the slag heap perimeter: 0 m (transect origin at the heap edge), 1, 5, 10, 50, and 100 m (see Figure 1). Two reference soil samples were collected to characterize the local background: one at a 300 m radius outside the immediate slag-impact zone (sample 267, hereafter Bg₃₀₀), and one at a 1.5 km radius from the heap boundary.

Four composite plant-tissue samples were collected adjacent to each cardinal-direction transect, comprising above-ground biomass of the two dominant native species—Centaurea pseudosquarrosa Mikheev ex Gabrieljan et Mikheev (Asteraceae) and Plantago lanceolata L. (Plantaginaceae)—pooled in approximately equal proportions to represent community-level uptake. Plants were sampled within 1–10 m of the slag-heap perimeter in each direction. This community-level composite design characterizes the bulk above-ground canopy burden but, by construction, does not resolve species-specific uptake; biological interpretation of the plant data is therefore restricted to descriptive bulk-canopy observations (Section 4.3 and Section 4.5). The dataset thus comprised 26 soil samples (24 transect points + 2 reference points) and 4 composite plant-tissue samples.

Immediately after collection, soil and plant samples were placed in pre-cleaned, double-bagged polyethylene containers, sealed, and individually labeled with sample identifier, date, GPS coordinates, transect direction, and distance from the slag-heap perimeter. Samples were transported from the field to the accredited Analytical Laboratory of LLP “KazEcoAnaliz” (Almaty, Kazakhstan) in insulated cooler boxes maintained at 4–8 °C using frozen ice packs to prevent microbial alteration of organic matter and minimize redistribution of mobile metal species between the sample and container surfaces during transit. Samples were transferred to the laboratory under a documented chain of custody and registered, weighed, and placed in temperature-controlled sample storage within 48 h of collection. Field handling, packaging, and transport procedures complied with the soil sampling requirements of GOST 17.4.4.02-2017 [24] and were consistent with the soil sample storage and transport guidance of ISO 18512:2007 [25].

All samples were air-dried at 20–25 °C in a dust-free, dark environment to constant mass (24–48 h), in accordance with the standard preparation protocol for total-element determination of plant and soil matrices by ICP-OES (ISO 17294-2; EPA Method 3050B) [26,27]. After drying, samples were sieved to <2 mm following removal of coarse debris and root material, and stored at +4 °C in sealed pre-cleaned containers pending acid digestion and analysis. The same air-drying protocol was applied to both soil and plant matrices to ensure analytical consistency; this approach is methodologically appropriate for total-metal determination and does not introduce volatilization losses at the applied temperature range, with the exception of trace mercury, for which a small loss cannot be excluded—however, mercury concentrations in plant tissue were below the analytical detection limit at all sampling locations (Section 3.3), so this constraint does not affect the reported results.

2.3. Chemical Analysis

Concentrations of As, Cd, Cr, Cu, Fe, Hg, Mn, Ni, Pb, and Zn in soil samples were determined by inductively coupled plasma optical emission spectrometry (ICP-OES; Shimadzu ICPE-9000, Kyoto, Japan) after acid digestion, in accordance with method M 02-902-157-10. Concentrations of the same elements in composite plant-tissue samples were determined according to method PND F 16.1:2.3:3.11-98 after wet mineralization. All analyses were performed during the period 27 April–5 May 2026 at the accredited Analytical Laboratory of LLP «KazEcoAnaliz» (accreditation certificate KZ.T.02.1017, issued 25 February 2026, valid for both methods and all ten target elements), Almaty, Kazakhstan, under Order N°23-26 (Test Protocols N°23-01 and N°23-02 dated 5 May 2026; reference background protocol for the 1.5 km sample dated 5 May 2026, analyzed in the same calibration batch). Analytical quality control was implemented in accordance with the laboratory’s ISO/IEC 17025-compliant internal procedures, including procedural blanks, analysis of certified reference materials within each batch with element-wise recoveries falling within the accreditation-defined acceptance limits (typically 80–120% for ICP-OES determinations), and replicate determinations on selected samples with relative standard deviations below 10% for major elements and below 20% for trace elements near the limit of quantification. Method detection limits were 0.2 mg·kg⁻¹ for As and 0.1 mg·kg⁻¹ for Hg in the plant matrix; values below detection were assigned half the detection limit for downstream calculations. Soil sample 260 (eastern transect, 100 m) returned matrix Fe (104 mg·kg⁻¹) and Zn (7.79 mg·kg⁻¹) values one to two orders of magnitude below those of adjacent samples on the same transect, most plausibly reflecting localized lithological heterogeneity (e.g., a quartz- or carbonate-dominated micro-substrate); this sample was retained in concentration means but excluded from Fe-normalized enrichment factor calculations to avoid artefactually inflated EF ratios.

2.4. Geochemical Pollution Indices

Six complementary indices were calculated to characterize single-element and aggregate contamination; their formulas, classification thresholds, and the reference values used in their calculation are consolidated in

Table 1. Pollution indices, classification thresholds, and regulatory reference values applied in this study.

Index/Reference	Formula or Values	Classification Thresholds	Source
Contamination factor (CF)	CF = Cn/Bn	<1 low; 1–3 moderate; 3–6 considerable; >6 very high	Hakanson [30]
Geoaccumulation index (Igeo)	Igeo = log2(Cn/1.5·Bn)	Class 0 (≤0) unpolluted → Class 6 (>5) extreme	Müller [31]
Enrichment factor (EF)	EF = (Cn/CFe)sample/(Bn/BFe)bg	<2 minimal; 2–5 moderate; 5–20 significant; 20–40 very high; >40 exclusively anthropogenic	Sutherland [32]
Pollution load index (PLI)	PLI = (CF1 × CF2 × … × CFn)^(1/n)	<1 unpolluted; >1 polluted	Tomlinson et al. [33]
Modified degree of contamination (mCd)	mCd = (ΣCFi)/n	<1.5 nil; 1.5–2 low; 2–4 moderate; 4–8 high; 8–16 very high; 16–32 extremely high; >32 ultra-high	Abrahim & Parker [34]
Potential ecological risk index (RI)	RI = ΣEri; Eri = Tri × CFi	<150 low; 150–300 moderate; 300–600 considerable; >600 very high	Hakanson [30]
Toxic response factors (Tri)	Pb = 5; Zn = 1; Cd = 30; Cu = 5; As = 10; Hg = 40; Cr = 2; Ni = 5	—	Hakanson [30]
Upper continental crust (UCC) values used as Bn (mg·kg−1)	Pb = 17; Zn = 67; Cd = 0.09; Cu = 28; As = 4.8; Hg = 0.05; Cr = 92; Ni = 47; Mn = 774; Fe = 39,200	Geochemical baseline	Rudnick & Gao [28]
Kazakhstan MPC for typical soils (mg·kg−1)	Pb = 130; Zn = 220; Cd = 1.0; Cu = 33; As = 2.0; Hg = 2.1; Cr = 100; Ni = 85	Regulatory exceedance threshold	Sanitary Rules, Republic of Kazakhstan [35]
REACH PNEC for soil (mg·kg−1)	Pb = 212; Zn = 123; Cd = 0.56; Cu = 65; As = 29; Hg = 0.22; Ni = 15; Cr = 21	RCR = PEC/PNEC; >1 unacceptable ecosystem risk	ECHA [29]

. The upper continental crust (UCC) compilation of Rudnick and Gao [28] was selected as the primary geochemical reference (Bn) over locally collected soils because both reference samples in the present study—at 300 m (Bg₃₀₀) and 1.5 km (Bg_1.5km)—showed elevated concentrations of at least one priority element (Section 3.1), and were therefore treated as distance-resolved exposure references rather than as a pristine geochemical baseline. Sample exceedance is reported against Kazakhstan’s maximum permissible concentrations (MPC) for typical soils (Table 1), and ecosystem risk in Section 2.6 against REACH-derived PNEC values for soil [29].

2.5. Vegetation Biomonitoring and Bioconcentration Factor

Above-ground biomass concentrations of each metal in the composite plant samples (Cplant) were related to the mean concentration of the same metal in soil along the corresponding transect (Csoil) by computing the ratio Cplant/Csoil. Because each composite plant sample combined approximately equal biomass proportions of two phylogenetically and physiologically distinct species (Centaurea pseudosquarrosa, Asteraceae, and Plantago lanceolata, Plantaginaceae), this ratio is reported here as a bulk vegetative canopy ratio—that is, as a descriptive index of the combined above-ground metal burden of the two species—rather than as a species-resolved Bioconcentration Factor. The conventional excluder/accumulator/hyperaccumulator classification of Baker [36] requires species-level measurements and cannot be applied to composite samples; it is therefore not invoked here. The major implications of this composite-sampling design for biological interpretation are addressed in Section 4.3 and Section 4.5.

2.6. Human-Health and Ecosystem Risk Assessment

Human health risk was assessed for an adult occupational worker scenario using the deterministic framework of the US EPA RAGS [37], with exposure factors taken from the US EPA Exposure Factors Handbook (2011 Edition) [38]. Three principal exposure pathways were integrated: incidental ingestion of soil, inhalation of resuspended particulates, and dermal contact. For each metal and each exposure pathway, the chronic daily intake (CDI, mg·kg⁻¹·d⁻¹) was calculated as:

CDI = (C × IR × EF × ED × CF)/(BW × AT),

(1)

where C is the metal concentration in soil (mg·kg⁻¹), IR is the pathway-specific intake rate (soil ingestion rate, inhalation rate, or dermal contact factor), EF is the exposure frequency (d·yr⁻¹), ED is the exposure duration (yr), CF is a unit-conversion factor, BW is the body weight (kg), and AT is the averaging time (d). Exposure-factor values followed the US EPA standard adult-worker defaults: body weight 70 kg, soil ingestion 50 mg·d⁻¹, inhalation rate 20 m³·⁻¹, exposed skin area 3300 cm², adherence factor 0.2 mg·cm⁻², particulate emission factor 1.36 × 10⁹ m³·kg⁻¹, exposure frequency 250 d·yr⁻¹, exposure duration 25 yr, and averaging time 70 yr (for carcinogenic endpoints) or 25 yr (for non-carcinogenic endpoints). The implications of using these defaults without site-specific localization are discussed in Section 4.5.

The carcinogenic risk (CR) for each metal–pathway combination was calculated by multiplying CDI by the slope factor (SF, (mg·kg⁻¹·d⁻¹)⁻¹) obtained from the US EPA IRIS database [39], accessed on 5 May 2026:

CR_i = CDI_i × SF_i,

(2)

Total carcinogenic risk was obtained by summing across metals and pathways. CR values below 10⁻⁶ are conventionally considered negligible; values between 10⁻⁶ and 10⁻⁴ are considered acceptable; and values above 10⁻⁴ are considered unacceptable [39].

Non-carcinogenic risk was assessed through the hazard quotient (HQ) and the hazard index (HI), using reference doses (RfD, mg·kg¹·d¹) from the US EPA IRIS database:

HQ_i = CDI_i/RfD_i

(3)

HI = Σ_i HQ_i,

(4)

HI values exceeding 1.0 indicate potential non-carcinogenic risk to the receptor.

Ecosystem risk was assessed via the risk characterization ratio (RCR), defined as the ratio of the measured site-mean soil concentration (predicted environmental concentration, PEC) to the element-specific predicted no-effect concentration in soil (PNEC):

RCR_i = PEC/PNEC_i,

(5)

The element-specific PNEC values for soil (mg·kg⁻¹ dry weight) were taken from REACH-registered substance dossiers compiled in the ECHA database [29]: Pb = 212, Zn = 123, Cd = 0.56, Cu = 65, As = 29, Hg = 0.22, Ni = 15, Cr = 21 (also reported in Table 1). The total ecosystem-risk score was computed as RCR_total = Σ_i RCR_i; RCR > 1 for any individual metal indicates an unacceptable risk to soil organisms for that metal at the site.

2.7. Statistical Analysis

Data are reported as mean ± standard deviation (SD). Differences in metal concentrations across cardinal-direction transects were assessed using one-way ANOVA with Tukey HSD post hoc (α = 0.05). Spatial decay of Pb along each transect was examined for monotonicity; given the observed non-monotonic pattern with secondary peaks at intermediate distances (Section 3.1), no single-source dispersion model was fitted. All calculations were performed in Python 3.11 using the pandas (version 2.0.3), NumPy (version 1.24.3), and SciPy (version 1.11.1) libraries. No inferential statistical test was performed for the soil-to-plant relationship: the composite-sampling design yielded only one plant sample per cardinal direction (n = 4), and rank-based or parametric correlations across such a small biologically heterogeneous dataset would carry no statistical power; the directional pattern in plant tissue concentrations is therefore reported as a descriptive observation only (Section 3.3 and Section 4.5).

3. Results

3.1. Heavy Metal Concentrations and Spatial Distribution

Soil concentrations of As, Cd, Cu, Hg, Pb, and Zn at the slag reprocessing site substantially exceeded both the upper continental crust geochemical background and Kazakhstan MPC values across all four cardinal-direction transects (Table 1, Figure 2). The site-mean ± SD concentrations across the 24 transect samples were 1754 ± 2447 mg·kg⁻¹ Pb, 1254 ± 873 mg·kg⁻¹ Zn, 53.7 ± 80 mg·kg⁻¹ Cd, 684 ± 886 mg·kg⁻¹ Cu, 1753 ± 2832 mg·kg⁻¹ As, and 8.0 ± 16 mg·kg⁻¹ Hg. The high standard deviations relative to the means reflect substantial directional and spatial heterogeneity rather than a smooth concentration gradient (see below).

The highest single-sample concentrations were observed at 5 m on the western transect (sample 263): 9350 mg·kg⁻¹ Pb (72× MPC), 290 mg·k⁻¹ Cd (290× MPC), 2870 mg·kg⁻¹ Zn (13× MPC), and 47.4 mg·kg⁻¹ Hg (23× MPC), all co-occurring in a single soil aliquot. Arsenic peaked at 5 m on the eastern transect (sample 257) at 10,900 mg·kg⁻¹, exceeding the Kazakhstan MPC by 5450-fold and forming a discrete hotspot likely associated with arsenopyrite-bearing slag fragments (Section 4.5). Copper reached 4230 mg·kg⁻¹ at 5 m on the southern transect (128× MPC), a finding not previously reported at this site. Chromium, nickel, and manganese remained at or below their respective MPCs at all transect points and showed CF and EF values consistent with a lithogenic origin (

Table 2. Mean ± SD and maximum-sample heavy-metal concentrations (mg·kg⁻¹) by cardinal-direction transect (n = 6 sampling distances per transect: 0, 1, 5, 10, 50, 100 m), reference samples, UCC, and Kazakhstan MPC. The “Maximum sample” row reports the absolute single-sample maximum measured anywhere at the site; the sampling location of each maximum is indicated in Section 3.1. UCC values from Rudnick and Gao [28].

Transect/Reference	Pb	Zn	Cd	Cu	As	Hg	Ni	Cr
South	1248 ± 466	1711 ± 808	27 ± 11	1387 ± 1377	366 ± 179	0.81 ± 0.81	17 ± 8	8 ± 4
North	735 ± 578	726 ± 481	14 ± 10	297 ± 195	289 ± 182	0.47 ± 0.16	9 ± 5	4 ± 2
East	258 ± 179	512 ± 301	8.5 ± 1.9	220 ± 72	2621 ± 3743	0.48 ± 0.30	6.7 ± 1.3	13 ± 13
West	4777 ± 3271	2066 ± 616	165 ± 94	832 ± 564	3735 ± 3043	30 ± 18	17 ± 6	12 ± 6
Site mean (n = 24)	1754 ± 2447	1254 ± 873	54 ± 80	684 ± 886	1753 ± 2832	8.0 ± 16	12 ± 7	9 ± 8
Bg300 (n = 1)	111	290	5.73	116	259	1.18	3.97	2.71
Bg1.5 km (n = 1)	205	79	0.30	37.6	4.8	<0.1	29.9	98.2
UCC [28]	17	67	0.09	28	4.8	0.05	47	92
Kazakhstan MPC	130	220	1.0	33	2.0	2.1	85	100
Maximum sample	9350	2870	290	4230	10,900	48	n.d. *	n.d. *
Max sample/MPC	72×	13×	290×	128×	5450×	23×	<1×	<1×

Note: * Indicates values below detection limit (n.d.). The symbol × indicates fold-excess of Kazakhstan Maximum Permissible Concentration (e.g., 72× = 72-fold the MPC). Sample 260 (East 100 m) returned anomalously low Fe and Zn and was retained in concentration means but excluded from EF calculations. MPC = maximum permissible concentration for typical soils, Republic of Kazakhstan; UCC = upper continental crust.

); these elements were therefore excluded from subsequent risk discussion.

The spatial distribution of metal concentrations was strongly asymmetric across cardinal directions, and the transect with the highest absolute single-sample peak was not always the transect with the highest mean concentration (Table 2). The western transect carried the highest mean values for Pb (4777 ± 3272 mg·kg⁻¹), Cd (165 ± 94), Hg (30 ± 18), and As (3735 ± 3044), and contained the absolute peak concentrations of Pb (9350 mg·kg⁻¹ at West 5 m, sample 263), Cd (290 at the same point), Zn (2870 at the same point), and Hg (47.4 at the same point), all four metals co-occurring in a single soil aliquot.

Arsenic showed a contrasting spatial pattern. The absolute peak (10,900 mg·kg⁻¹ at East 5 m, sample 257) occurred on the eastern transect, but the mean As concentration on the eastern transect (2621 ± 3743 mg·kg⁻¹) was lower than that on the western transect (3735 ± 3044 mg·kg⁻¹), because the eastern signal is dominated by a single extreme hotspot (10,900 mg·kg⁻¹) flanked by markedly lower concentrations at the other five eastern sampling distances (630, 2010, 1113, 194, and 879 mg·kg⁻¹ at 0, 1, 10, 50, and 100 m respectively), whereas the western transect exhibits broadly elevated As across multiple distances (3890 at 0 m, 7610 at 50 m, 7900 at 100 m), without any single extreme outlier. This pattern is consistent with two different contamination processes operating on the two transects: a discrete arsenopyrite-bearing slag fragment producing a localized point hotspot at East 5 m, versus broadly distributed atmospheric deposition and surface-runoff redistribution producing diffusely elevated As across the entire western transect.

Copper showed a third distribution pattern: although its mean was highest at the southern transect (1387 ± 1508 mg·kg⁻¹), the absolute peak (4230 mg·kg⁻¹ at South 5 m, sample 245) was substantially higher than any value recorded on the western transect, indicating that Cu, like As, exhibits localized hotspot behavior rather than smooth radial decay. Pb concentrations did not decline monotonically with distance from the slag heap on either the western or northern transects: a secondary Pb maximum of 7850 mg·kg⁻¹ was recorded at 100 m on the western transect, demonstrating that a simple radial dispersion model does not adequately describe the contamination plume around the heap. Together, these directional asymmetries—different mean-peak relationships across metals, secondary off-perimeter maxima, and the co-occurrence of multiple priority metals in a single West-transect aliquot—are consistent with patchy deposition controlled by the prevailing easterly–westerly wind regime, by topographic gradients, and by the discrete redistribution of arsenopyrite- and galena-bearing slag fragments during episodic surface-runoff events.

The two background reference samples revealed an instructive contrast (Table 2). The 300 m sample (Bg300) still showed clearly elevated As (259 mg·kg⁻¹), Cd (5.73), Hg (1.18), and Cu (116)—values 30–60× the UCC background—confirming that the impact zone of this site extends beyond 300 m. The 1.5 km sample (Bg1.5 km), in contrast, returned As, Cd, and Hg values close to UCC; however, its Pb concentration of 205 mg·kg⁻¹ (12× UCC) indicates that aerial Pb deposition from the heap, or from related historical operations within the Achisai–Shymkent corridor, extends to at least 1.5 km. For this reason, all pollution indices reported here use UCC as the geochemical reference; the local 1.5 km sample is reported alongside for context but is not used as Bn in CF and Igeo calculations.

The persistence of elevated As, Cd, Hg, and Cu at 300 m and of elevated Pb at 1.5 km most plausibly reflects three superimposed processes acting on overlapping timescales. First, long-range aeolian deposition of fine slag particulates (predominantly <10 μm aerodynamic diameter) accumulated during the multi-decadal operational period of the former Shymkent lead plant (c. 1940–2012) under the prevailing easterly and westerly wind regime [8]; particles in this size class can remain airborne over kilometre-scale distances under semi-arid conditions and account for the broad spatial footprint of the contamination signal documented at the 300 m reference point. Second, the site lies within the broader Achisai–Shymkent metallurgical corridor, which has hosted Pb–Zn extraction and processing since the early 20th century; the elevated Pb concentration of 205 mg·kg⁻¹ at the 1.5 km reference point (12× UCC), in contrast to the near-UCC Cd, As, and Hg values at the same point, is consistent with a cumulative regional Pb-deposition signal that cannot be attributed to the present slag heap alone [9]. Third, episodic high-intensity precipitation events generate surface runoff that redistributes coarser arsenopyrite- and galena-bearing slag fragments along topographic gradients toward downslope catchment areas, producing localized off-perimeter hotspots inconsistent with a smooth radial decay model. Together, these three mechanisms—atmospheric dispersion, cumulative regional emissions, and runoff-mediated particulate redistribution—account for the observed extension of the impact zone well beyond 300 m and provide the methodological rationale for the use of UCC, rather than locally collected reference soils, as the geochemical baseline (Bn) in all index calculations reported below.

3.2. Geochemical Pollution Indices

All six pollution indices placed the site in the highest contamination categories for Pb, Cd, Zn, As, and Hg, while confirming that Cr, Ni, and Mn were within the lithogenic range (

Table 3. Geochemical pollution indices for individual heavy metals at the Shymkent Pb–Zn slag reprocessing site, referenced to the upper continental crust [28]. Values are mean (max) across 24 transect samples; EF based on n = 23 after exclusion of sample 260.

Metal	CF Mean (Max)	Igeo * Mean (Max)	EF Mean (Max)	Igeo Cl. 6 (n/24)	Classification
Pb	103 (550)	4.93 (8.52)	2369 (15,511)	14	Extreme
Zn	19 (43)	3.03 (4.84)	391 (1208)	0	Heavily contaminated
Cd	597 (3222)	7.37 (11.07)	11,501 (90,871)	22	Extreme
Cu	24 (151)	3.27 (6.65)	526 (3117)	3	Extreme
As	365 (2271)	6.51 (10.56)	5066 (30,008)	18	Extreme
Hg	160 (948)	3.99 (9.30)	3654 (26,735)	6	Extreme
Cr	0.10 (0.40)	−4.33 (−1.90)	1.7 (6.4)	0	Lithogenic
Ni	0.26 (0.58)	−2.79 (−1.38)	5.2 (16.3)	0	Lithogenic

CF—contamination factor; Igeo *—geoaccumulation index; EF—Fe-normalized enrichment factor. EF > 40 indicates exclusively anthropogenic origin [34]. Aggregate indices: PLI mean = 9.6 (max 35.5); mCd mean = 141 (max 652); RI mean = 28,605 (max 138,355). See Section 3.2 for classifications and interpretations.

). Mean CF values were Cd = 597, As = 365, Hg = 160, Pb = 103, Cu = 24, and Zn = 19—all greatly exceeding the threshold of 6 for very high contamination [13,31]. Mean Igeo values placed Cd (7.4), As (6.5), Pb (4.9), Hg (4.0), and Zn (3.0) in the heavily-to-extremely contaminated range; the corresponding maxima (Cd 11.1, As 10.6, Hg 9.3, Pb 8.5, Cu 6.7) all reached Igeo Class 6 (extreme contamination, Igeo > 5). Of the 24 transect samples, Class 6 was reached by 22 samples for Cd, 18 for As, 14 for Pb, 6 for Hg, and 3 for Cu.

Enrichment factors (Fe-normalized) far exceeded the conventional threshold of 40, taken to indicate an exclusively anthropogenic origin [34]: mean EF values reached 11,501 for Cd, 5066 for As, 3654 for Hg, 2369 for Pb, and 526 for Cu, with sample-level maxima of 90,871 (Cd), 30,008 (As), and 26,735 (Hg). EF values for Cr (mean 1.7), Ni (5.2), and Mn (5.9) remained below 10, confirming a lithogenic origin for these elements and supporting the choice of Fe as the normalizing element. The combined indices PLI (mean 9.6, max 35.5), mCd (mean 141, max 652), and RI (mean 28,605, max 138,355) all classified the site as ultra-high multi-element contamination, with RI exceeding the conventional very-high threshold of 600 by approximately two orders of magnitude. Among individual element contributions to the Hakanson RI, mercury and cadmium dominated (mean Er for Hg = 6396; for Cd = 5370; for As = 3652) owing to their high toxic-response factors of 40 and 30, respectively [30]; lead, despite the highest absolute concentrations, contributed proportionally less (mean Er = 516).

The fold-excess values of all indices relative to Kazakhstan MPC are presented in

Table 4. Heavy-metal concentrations (mg·kg⁻¹ dry weight) in composite above-ground plant tissue at the four cardinal-direction transects, listed in order of increasing mean soil Cd concentration. Each composite comprised approximately equal biomass of Centaurea pseudosquarrosa and Plantago lanceolata.

Transect	Soil Cd Mean (mg·kg−1)	Pb (Plant)	Zn (Plant)	Cd (Plant)	Cu (Plant)	As (Plant)	Hg (Plant)
East	8.5	93.7	1120	44.8	337	<0.2	<0.1
North	14.0	221	1080	38.8	300	<0.2	<0.1
South	26.7	134	426	17.5	156	<0.2	<0.1
West	165	206	932	18.6	288	<0.2	<0.1

All plant concentrations in mg·kg⁻¹ dry weight. As <0.2 and Hg < 0.1 mg·kg⁻¹ are method detection limits in the plant matrix, values are reported as half the detection limit for downstream calculations.

.

3.3. Vegetation Tissue Analysis and Bulk Canopy Ratios

The bulk vegetative canopy ratio (C_plant/C_soil-mean) calculated for each cardinal-direction composite is reported in

Table 5. Bulk vegetative canopy ratios (C_plant/C_soil-mean) for composite plant tissue (combined Centaurea pseudosquarrosa and Plantago lanceolata, approximately equal biomass proportions) at the four cardinal-direction transects. The ratios are reported as a descriptive index of combined above-ground metal burden across the two species, not as a species-resolved Bioconcentration Factor; uptake-strategy classification is not invoked here because composite samples do not permit species-level physiological interpretation (see Section 4.3).

Transect	BCF Pb	BCF Zn	BCF Cd	BCF Cu
East	0.36	2.19	5.24	1.53
North	0.30	1.49	2.76	1.01
South	0.11	0.25	0.66	0.11
West	0.04	0.45	0.11	0.35

Ratios are unitless. As and Hg were excluded from the ratio calculation because plant concentrations were below detection limits at all four directional composites (Table 4). The Baker [36] excluder/accumulator/hyperaccumulator framework requires species-level measurements and is not applied to these composite values.

.

The ratio varied substantially across directions: on the eastern transect, where mean soil contamination was moderate, the ratios for Cd (5.24), Zn (2.19), and Cu (1.53) exceeded 1, indicating that the combined above-ground tissue burden of the two species exceeded the corresponding mean soil concentration. On the northern transect, the ratios for Cd (2.76) and Zn (1.49) also exceeded 1, while at the southern and western transects, where mean soil contamination was intermediate to extreme, all ratios were below 1 (e.g., Cd 0.66 south, 0.11 west; Pb 0.11 south, 0.04 west). Across the four directional composites, the bulk canopy ratio for Cd, Pb, and Zn decreased visually with increasing mean soil contamination.

These ratios are reported here as descriptive bulk-canopy observations only. They do not support inference about the metal-uptake strategy of either constituent species, because each composite combined approximately equal biomass proportions of Centaurea pseudosquarrosa (Asteraceae) and Plantago lanceolata (Plantaginaceae), and the resulting signal averages potentially divergent species-level uptake, translocation, and detoxification patterns into a single value (Section 4.3). For the same reason, no inferential statistical test (rank-based or parametric correlation) is reported across these four composite values: with a single observation per cardinal direction and biologically heterogeneous composites, such tests would have negligible statistical power and would not constitute meaningful evidence for a soil-to-plant relationship. The directional pattern is noted here as a visual observation that motivates expanded, species-resolved, replicated sampling in the next phase of the project (Section 4.5).

3.4. Human-Health and Ecosystem Risk

Health and ecosystem risk metrics for the adult occupational scenario are summarized in

Table 6. Summary of human-health and ecosystem risk at the Shymkent slag reprocessing site. Worst-case = sampling point with highest single-metal concentration (As: East 5 m, sample 257; other metals: West 5 m, sample 263). Site mean = mean across 24 transect samples.

Risk Metric	Worst Case	Site Mean	Threshold	Interpretation
Carcinogenic risk (CR)	4.3 × 10−3	7.0 × 10−4	10−6 negligible; 10−4 acceptable	Unacceptable
Non-carcinogenic hazard index (HI)	26.6	4.3	1.0	Adverse effects expected
RCRtotal	—	223	1.0	Critical ecosystem risk
RCRCd	—	96	1.0	Dominant ecological driver
RCRAs	—	60	1.0	Critical
RCRHg	—	36	1.0	Critical
RCRCu	—	11	1.0	Very high
RCRZn	—	10	1.0	Very high
RCRPb	—	8.3	1.0	High

. The total carcinogenic risk at the worst-case sampling point reached CR = 4.3 × 10⁻³, exceeding the US EPA acceptable threshold of 10⁻⁴ by 43-fold and the negligible threshold of 10⁻⁶ by 4310-fold. Arsenic was the dominant contributor (92.7% of CR), followed by cadmium (7.3%); chromium and nickel together contributed less than 0.01%. Across the 24 transect samples, the site-mean carcinogenic risk was CR = 7.0 × 10⁻⁴, still 7-fold above the US EPA acceptable limit. The non-carcinogenic hazard index reached HI = 26.6 at the worst-case sampling point (As-driven, 93%; Pb 5%; Cd 1%) and HI = 4.3 at the site mean—both far above the threshold of 1, indicating adverse non-carcinogenic effects.

Ecosystem risk, based on PEC/PNEC ratios, reached RCR_total = 223 at the site mean, with cadmium as the dominant component (RCR_Cd = 96, 43% of total), followed by arsenic (RCR_As = 60, 27%), mercury (RCR_Hg = 36, 16%), copper (RCR_Cu = 11), zinc (RCR_Zn = 10), and lead (RCR_Pb = 8). Although Pb dominated the absolute mass loading and the individual ecological risk index in the Hakanson framework (mean Er_Pb = 516), Cd emerged as the dominant ecological-risk driver in the PEC/PNEC framework, owing to its extremely low PNEC (0.56 mg·kg⁻¹) and high bioavailability under the site’s neutral-to-alkaline soil conditions (pH 7.5–8.5). Arsenic contributed comparably high ecosystem risk in absolute terms—a finding not captured by the Hakanson framework alone and one that motivates the dual As + Cd priority-contaminant designation argued in Section 4.

4. Discussion

The geochemical, plant-tissue, and risk evidence presented above place the Shymkent Pb–Zn slag reprocessing site among the most severely contaminated metallurgical legacy sites documented in Central Asia and globally. In the following sections, we develop the implications of this evidence for contaminant prioritization, vegetation-mediated containment, and site management.

4.1. Contamination Context

Peak soil concentrations at Shymkent (Pb 9350, Zn 2870, Cd 290, Cu 4230, As 10,900, Hg 47.4 mg·kg⁻¹) substantially exceed both Kazakhstan MPC values and concentrations reported for analogous Pb–Zn smelter-affected soils on three continents (

Table 7. Comparison of peak concentrations (mg·kg⁻¹) of priority metals in soils at the Shymkent slag reprocessing site with published values from representative Pb–Zn smelter-affected soils worldwide.

Site	Country	Pb	Zn	Cd	As	Source
Shymkent slag site	Kazakhstan	9350	2870	290	10,900	Present study
Henan Pb–Zn smelter	China	up to ~8000	1526–6812	4.5–28.3	n.d.	[15]
Příbram smelter	Czech Republic	1200–5600	410–2800	3.2–18	20–80	[16]
Kentau–Aksu corridor	Kazakhstan	300–1500	180–620	0.4–1.65	n.d.	[17]
Middle Spiš Pb–Zn–Cu–Hg mining area	Slovakia	340–1850	290–960	1.2–6.8	n.d.	[18]
Kabwe mine/smelter	Zambia	780–14,100	n.d.	2.1–48	n.d.	[9]

n.d. = not determined/not reported in the cited study.

). Among published comparator sites, only the Kabwe mine–smelter complex in Zambia approaches the Pb maximum recorded here, whereas the Cd, Cu, As, and Hg maxima at Shymkent exceed all listed comparators by factors of 4–100×. Notably, the 10,900 mg·kg⁻¹ maximum at the eastern transect 5 m point is unprecedented at Pb–Zn smelter sites globally and is a defining feature of this site, warranting its independent recognition as an As-priority contamination zone. This designation is supported by three additional lines of evidence within our dataset. First, the high Igeo class 6 frequency for As (18 of 24 transect samples; Table 3) shows that extreme As contamination is not confined to the 10,900 mg·kg⁻¹ hotspot but extends across most of the impact zone. Second, the As enrichment factor (mean 5066; maximum 30,008; Table 3) exceeds the conventional threshold for exclusive anthropogenic origin [34] by two to three orders of magnitude, confirming that the As burden cannot be attributed to natural lithogenic variability [28]. Third, As dominates the site’s carcinogenic risk profile (92.7% of CR at the worst-case point and the dominant contributor to site-mean CR; Table 6), elevating the human-health relevance of the As signal above that of any other detected metal at this site. Together with the unprecedented maximum concentration, these spatial, geochemical, and toxicological lines of evidence justify the independent recognition of this site as an As-priority contamination zone.

The hierarchy EF_Cd > EF_As > EF_Hg > EFPb > EF_Zn differs from the EF_Pb > EF_Zn > EF_Cd ranking commonly reported at Western European and Chinese Pb–Zn smelter sites [15,16], reflecting the unusually heavy contribution of Cd and As to the Shymkent contamination signature. This is consistent with the polymetallic ore feed of the Achisai deposit, in which arsenopyrite, sphalerite, and Cd-bearing sulfides constitute a larger proportion of the original ore than at most monometallic Pb–Zn operations.

4.2. Arsenic and Cadmium as Co-Priority Contaminants

The risk-assessment results (Section 3.4) reveal a striking disconnect between mass-based and toxicity-based contaminant prioritization at this site [37]. By mass, Pb dominates the soil burden at most sampling points and is the contaminant most readily addressed by conventional regulatory frameworks. However, the human-health risk assessment identifies arsenic as the dominant driver of carcinogenic and non-carcinogenic risk (92.7% of CR; 93% of HI at the worst-case point), reflecting both the extreme As hotspot of 10,900 mg·kg⁻¹ on the eastern transect and the high IRIS slope factor for inorganic As (SFinh = 15.1 (mg·kg⁻¹·d⁻¹)⁻¹). The ecosystem risk assessment, in turn, identifies cadmium as the dominant driver (RCR_Cd = 96, 43% of the total) owing to its extremely low PNEC (0.56 mg·kg⁻¹) and high bioavailability under the neutral-to-alkaline soil conditions documented at this site (pH 7.5–8.5).

Three structural causes explain the disconnect between mass-based and toxicity-based prioritization and clarify why the contaminant with the largest absolute soil burden need not be the dominant driver of risk at the same site. First, the IRIS inhalation slope factor for inorganic arsenic (SFinh = 15.1 (mg·kg⁻¹·d⁻¹)⁻¹) is approximately three orders of magnitude higher than that of lead and four orders of magnitude higher than that of zinc, so that [37] even a moderately elevated As concentration produces a far larger contribution to CR than a much higher Pb concentration at the same point. Second, the REACH-derived PNEC for soil cadmium (0.56 mg·kg⁻¹) is the lowest among the eight priority metals assessed here—roughly 400× lower than the PNEC for Pb (212 mg·kg⁻¹) [29]—so that even modest Cd enrichment produces a disproportionately large RCR_Cd in the ecosystem risk framework. Third, the neutral-to-alkaline soil conditions documented at this site (pH 7.5–8.5) preferentially mobilize Cd²⁺ relative to Pb²⁺, because Pb²⁺ is strongly retained by carbonate and Fe-oxide surfaces under these conditions, whereas Cd²⁺ is comparatively more bioavailable to plant roots and soil organisms [40]. The combined effect of these three factors—differential toxicological potency, differential PNEC stringency, and differential pH-dependent bioavailability—is that absolute mass loading and toxicity-weighted risk identify different priority contaminants at the same site, even when measured concentrations and risk frameworks are both internally consistent.

These three lines of evidence—mass loading, human-health risk, and ecosystem risk—therefore identify three priority contaminants (Pb, As, and Cd, respectively). A defensible management framework for this site must address all three simultaneously rather than treating Pb in isolation. Approaches that reduce Pb mobility without addressing Cd or As—such as phosphate amendments that form stable pyromorphite [Pb₅(PO₄)₃Cl] but leave Cd and As in exchangeable form—would fail to reduce the principal ecosystem and human-health risk drivers. Conversely, a strategy targeting only the As hotspot would overlook the broadly distributed Cd and Pb burden across the impact zone. The implication is clear: Pb–Zn smelter legacy sites should not be classified or prioritized based on Pb concentration alone, particularly at sites with polymetallic ore feeds.

4.3. Vegetation Biomonitoring: Bulk Canopy Metal Burden

Composite aboveground plant tissue (a mixture of Centaurea pseudosquarrosa and Plantago lanceolata, with approximately equal biomass proportions) was analyzed along the four cardinal-direction transects to characterize the bulk metal burden carried by the dominant native vegetation canopy at the site. Tissue Cd concentrations ranged from 17.5 mg·kg⁻¹ DW (south) to 44.8 mg·kg⁻¹ DW (east); tissue Zn from 426 to 1120 mg·kg⁻¹ DW; and tissue Pb from 93.7 to 221 mg·kg⁻¹ DW (Table 4). Arsenic and mercury concentrations in plant tissue were below detection limits in all four directions, despite very high soil concentrations of these elements on the western and eastern transects.

The bulk canopy ratio Cplant/Csoil (Table 5) exceeded 1 for several metals at the eastern and northern transects (Cd 5.24 and 2.76; Zn 2.19 and 1.49) and was consistently below 1 across all metals at the southern and western transects, where soil contamination was most severe. These ratios reflect the combined aboveground burden of the two species; they do not allow either species to be classified as an excluder or an accumulator in the sense of Baker [36], because the underlying composite samples conflate species-specific uptake, translocation, and detoxification patterns into a single signal. C. pseudosquarrosa (Asteraceae) and P. lanceolata (Plantaginaceae) belong to phylogenetically distinct families that can exhibit opposite uptake strategies at the same site [14,41,42]; the composite design therefore averages potentially divergent species-level behaviors into an undifferentiated bulk signal that lacks a straightforward physiological interpretation.

For the same reason, the apparent inverse rank ordering between mean soil Cd, Pb, or Zn and the corresponding bulk canopy ratio across the four directional composites is interpreted here as a descriptive observation only. No mechanistic inference about physiological down-regulation of root uptake, selective mortality of accumulator phenotypes, or rhizosphere biogeochemical immobilization can be drawn from these four data points: each direction yielded a single composite sample combining two species, and species-resolved tissue concentrations, paired root–shoot transfer factors, and rhizosphere chemistry would all be required to support any such interpretation. The data presented here, therefore, document the combined metal burden in the dominant above-ground vegetation at this site but do not provide evidence of the metal-uptake strategy of either constituent species.

Species-resolved tissue analysis, with paired root and shoot sampling of C. pseudosquarrosa and P. lanceolata analyzed independently and replicated within each cardinal direction, is mandatory before any uptake-strategy interpretation can be supported. This species-resolved campaign is scheduled within the 2026–2028 continuation of the AP25795537 project (Section 4.6). Pending those data, no recommendation regarding the phytostabilization or phytoremediation potential of either species can be drawn from the present bulk-canopy results, and any such recommendation in the present manuscript has been removed.

4.4. Implications for Site Management

Integrating geochemical severity, plant biomonitoring, and risk-assessment data supports a preliminary three-tier zonation for remediation prioritization. Framed as a sustainable management strategy, this zonation is designed to reconcile two objectives often treated separately at legacy sites: recovering residual value from the slag as a secondary raw material and arresting the off-site migration of contaminants that threaten adjacent agricultural land, the Badam River, and the residential districts of Shymkent. The three tiers are:

Zone 1 (0–10 m perimeter; PLI > 20; RCR_total > 100): Active engineered intervention is essential. Options include (i) physical containment of the slag heap with a multilayer cover and surface stabilization; (ii) selective pyrometallurgical reprocessing of the slag, which has been shown to recover >70% of residual Zn and Pb from comparable feedstock while fixing Cd in stable gangue phases [4,5]; and (iii) hotspot excavation and off-site treatment of the highest As-bearing zones (e.g., East 5 m). Selective pyrometallurgical reprocessing, in particular, not only reduces the residual metal burden at this site but also offers a circular-economy pathway for resource recovery from secondary metallurgical materials, simultaneously addressing contamination and resource-recovery opportunities. The Badam River, located 20 m west of the facility, makes runoff control a particular priority for this zone; direct quantification of fluvial metal transport through dedicated sediment sampling is scheduled for the next phase of the project (Section 4.6).

Zone 2 (10–100 m; PLI 5–20; RCR_total 10–100): Phytostabilization using C. pseudosquarrosa and P. lanceolata, potentially combined with mycorrhizal inoculation and biochar amendment, is a cost-effective intervention pending the confirmatory studies outlined in Section 4.3. Concurrent monitoring of plant tissue concentrations, soil pH, and DTPA-extractable Cd should accompany any pilot-scale deployment.

Zone 3 (>100 m; PLI < 5; RCR_total < 10): Monitored natural attenuation with periodic resampling is appropriate, with priority attention to the west, where elevated concentrations were detected at 100 m and the 1.5 km background sample also showed elevated Pb. The persistence of measurable Pb deposition at 1.5 km has direct implications for residential exposure risk in the western and southwestern districts of Shymkent and warrants integration with city-level air-quality monitoring data.

4.5. Limitations

Several limitations of this study should be acknowledged. The most important is the composite vegetation-sampling design: as noted in Section 2.2, each directional plant sample pooled two phylogenetically distinct species—Centaurea pseudosquarrosa (Asteraceae) and Plantago lanceolata (Plantaginaceae)—in approximately equal proportions, so the resulting C_plant/C_soil ratios reflect a bulk canopy burden and cannot be interpreted in terms of species-level uptake strategy or the excluder/accumulator categories of Baker [36]; the four-direction design also precludes inferential statistics on the plant data (Section 3.3). Second, the study rests on a single April 2026 topsoil (0–50 cm) campaign and therefore does not capture seasonal variability in metal mobility under the episodic moisture regime of the semi-arid Turkestan Region; in particular, the As hotspot at sample 257 (East 5 m, 10,900 mg·kg⁻¹) is likely associated with discrete arsenopyrite-bearing slag fragments rather than a smooth concentration gradient, which finer-resolution sampling along the eastern transect would resolve. Third, the human-health risk estimates (Section 2.6) are deterministic single-point values calculated from US EPA RAGS default exposure factors for an adult occupational scenario, without adjustment for site-specific conditions; given the arid, high-wind, dust-prone climate of Shymkent, the reported worst-case CR of 4.3 × 10⁻³ and HI of 26.6 should be regarded as conservative screening estimates rather than best-estimate central tendencies. These limitations directly define the priority directions for follow-up research set out in Section 4.6.

4.6. Future Research Directions

Priority directions for follow-up research are outlined in the 2026–2028 continuation of the AP25795537 research project and comprise five interlinked work packages. First, seasonal resampling of the impact zone is planned for spring, summer, and autumn 2026–2027 to capture temporal variability in metal mobility, dust deposition, and surface runoff under contrasting moisture regimes. Second, BCR sequential extraction [43] is scheduled for the 2027 analytical campaign on archived and freshly collected soil samples to determine operationally defined speciation of Pb, Zn, Cd, and As and to refine bioavailability estimates beyond the total-concentration measurements reported here. Third, species-resolved tissue analysis with paired root and shoot sampling of Centaurea pseudosquarrosa and Plantago lanceolata is planned for the 2026–2027 vegetation campaign to compute species-level transfer factors, distinguish the physiological and ecological mechanisms underlying the community-level accumulator-to-excluder shift documented in this study, and confirm or refine the preliminary phytostabilizer designation. Fourth, a dedicated Badam River sediment-sampling campaign is scheduled for August–September 2026. Fifth, a probabilistic Monte Carlo re-analysis of the human-health risk component is planned for the 2027 modeling campaign, incorporating site-specific occupational exposure parameters (localized inhalation, ingestion, and dermal adherence distributions for outdoor industrial workers in semi-arid urban settings) and a seasonal wind-blown dust factor; this analysis will replace the deterministic CR and HI point estimates reported here with full probability distributions, allowing direct evaluation of exceedance probability at the US EPA 10⁻⁴ and 10⁻⁶ thresholds. Together, these scheduled activities will deliver the species-, season-, speciation-, and uncertainty-resolved evidence base required to support the operational deployment of the three-tier remediation framework proposed here.

Complete methodological details, supplementary data tables, and additional analytical results, including per-sample heavy-metal concentrations, computed pollution indices, and risk metrics for all sampled locations, are provided in the Supplementary Materials.

5. Conclusions

Integrated geochemical, vegetation, and risk assessment at an active Pb–Zn slag reprocessing site in Shymkent (Southern Kazakhstan) places it among the most severely contaminated Pb–Zn legacy sites documented in Central Asia and globally: cadmium and arsenic reached Geoaccumulation Class 6 across most of the impact zone, and enrichment factors for all six priority metals exceeded the anthropogenic-origin threshold by two to three orders of magnitude.

The principal contribution of this work is the identification of three distinct priority contaminants, depending on the assessment framework—lead by absolute mass loading, arsenic by human health risk (at worst-case points, >90% of the carcinogenic risk and hazard index), and cadmium by ecosystem risk. This disconnect between mass-based and toxicity-based prioritization argues for elevating arsenic and cadmium to co-priority status alongside lead at polymetallic Pb–Zn legacy sites, rather than targeting Pb in isolation. Composite vegetation analysis documented the bulk above-ground canopy burden but, by design, does not support species-level uptake inference; therefore, species-resolved tissue analysis is identified as a mandatory next step before any phytoremediation recommendation can be made (Section 4.5 and Section 4.6).

The integrated framework deployed here is directly transferable to analogous Pb–Zn reprocessing legacy sites and supports a three-tier zonation for remediation prioritization: active engineered intervention within the 0–10 m perimeter, pilot interventions and further characterization in the 10–100 m zone, and monitored attenuation beyond 100 m. By coupling secondary resource recovery with contamination control, this evidence-based framework advances the sustainable rehabilitation of metallurgical legacy land and contributes to UN Sustainable Development Goals 3, 11, 12, and 15.