Reclaiming Mercury Tailings as Urban Parks: Evidence from Soil and Vegetation Responses

Abstract

The switch in land use of abandoned tailings can precondition their reuse as newly built parks. This study investigated the feasibility of reusing a remediated mercury (Hg) retorting site in Wanshan, Guizhou Province, China, as a functional urban park by assessing residual heavy metal risks and associated vegetation responses. Field investigations were conducted across 31 park sites distributed along an east–west geographical gradient from the former mining area to urban parks, using replicated plots to sample the surface soils and dominant plant species. The concentrations of arsenic (As), cadmium (Cd), mercury (Hg), manganese (Mn), and lead (Pb) in soil and plant tissues were quantified using inductively coupled plasma–mass spectrometry, and vegetation structure and diversity were evaluated using standard community indices. The results showed significant spatial variability in soil and plant metal concentrations, with higher levels generally observed near historically impacted areas of the mine. However, all soil metal concentrations were below the national safety thresholds. Plant tissues exhibit controlled metal accumulation within normal or regulated ranges, reflecting the effective screening of tolerant and hyperaccumulating species. Increasing heavy metal concentrations were associated with reduced vegetation coverage, height, and diversity along the gradient. Overall, the findings indicate that the reclaimed Hg retorting site almost met ecological safety requirements, but more data on deep soils, groundwater, and long-term observations are needed to draw more conclusive conclusions.

1. Introduction

Mining tailing sites are often abandoned after resource extraction; however, they still hold significant potential for repurposing as urban parks, as they retain green spaces for recreation and biodiversity restoration in densely populated areas [1,2]. Such transformations can enhance urban ecosystems by converting degraded lands into functional landscapes and promoting environmental sustainability and community well-being [3]. However, a major obstacle to this reuse is severe soil contamination by heavy metals, including As, Cd, Hg, Mn and Pb. These pose threats to regional environments and human health [4,5]. For instance, As and Cd often exceed safe thresholds in tailings; hence, they have been reported to lead to bioaccumulation in food chains and potential toxicity [6]. Similarly, the persistence of Hg and Pb in soils can inhibit vegetation establishment, exacerbating erosion and leaching into water sources [7,8,9]. Although less studied, Mn overload alters soil nutrient dynamics and stresses the native flora [10]. Overcoming these barriers requires effective heavy metal extraction to render the soil safe for park development [11]. Phytoremediation utilizes hyperaccumulating plants to uptake and sequester these metals and has emerged as a viable and low-cost strategy for site cleanup [12,13]. Once metals are removed, rehabilitated tailings can support diverse plant communities and further facilitate the establishment of secure urban parks [14].

Soil contamination by heavy metals is a widespread environmental issue arising from geogenic processes and intensive anthropogenic activities. Soil As is commonly derived from sulfide mineral weathering and hydrothermal deposits; however, its enrichment is strongly driven by gold mining, smelting, and long-term use of As-based pesticides, resulting in elevated bioavailable levels in mining-affected soils [15]. Cd is typically introduced through non-ferrous metal smelting, phosphate fertilizers, and industrial emissions; hence, it readily accumulates in agricultural topsoil because of its high mobility and weak sorption [16,17]. Mercury contamination is closely associated with cinnabar mining, coal combustion, and atmospheric deposition, with legacy Hg mines acting as long-term sources of soil Hg, which is subsequently transferred to aquatic and terrestrial biota [18]. Manganese can be an essential micronutrient for some species, but it can also easily reach toxic concentrations in soils affected by mining and metallurgical activities, where altered redox conditions enhance Mn solubility and plant uptake [19,20]. Lead primarily originates from ore mining, smelting, traffic emissions, and industrial waste and persists in soils because of its strong binding to mineral and organic phases, resulting in long-term exposure risk [17,21]. The environmental hazards of heavy metals are manifested through their persistence, bioaccumulation and toxicity across trophic levels. As and Cd pose severe risks to soil–plant systems by impairing plant growth and facilitating their transfer into food chains [22]. Mercury is of particular concern because of its methylation and biomagnification potential, which threatens ecosystems and human health, even at low concentrations [18]. Excess Mn disrupts photosynthesis and nutrient balance in plants, whereas chronic Pb exposure reduces soil microbial activity and biodiversity [16,20]. Collectively, these metals degrade soil functions and ecosystem services, highlighting the urgency of understanding their sources and impacts to support effective risk assessment and remediation strategies.

Karst soils develop on carbonate bedrock and are characterized by thin profiles, high rock–soil interaction, and strong geochemical heterogeneity, which favor the accumulation and mobility of toxic elements. Arsenic is frequently enriched in karst regions affected by mining or natural mineralization, where its spatial variability reflects both the parent material and hydrological transport through fissures and sinkholes [23,24]. Cd is often present at low background levels, but it can still reach ecotoxic thresholds in karst soils due to atmospheric deposition and legacy mining, which poses risks to crops and soil biota [17,25]. Mercury is of particular concern in karst ecosystems hosting historical Hg mining, such as the Wanshan area in southwest China [26]. Karst soil–rock systems can act as long-term Hg reservoirs, whereas episodic leaching and erosion facilitate its redistribution to surrounding soils and waters, amplifying ecological exposure [9,18,27]. Mn is an essential micronutrient for Karst plants and is highly reactive in Karst soils. Excessive Mn accumulation may occur under acidic or redox-fluctuating conditions, leading to phytotoxicity and altered plant–microbe interactions [19,20]. Lead commonly co-occurs with Zn and Cd in karst mining landscapes, and its strong affinity for soil particles does not preclude biological uptake, which enables its transfer to vegetation and food webs [15,28]. Taken together, elevated As, Cd, Hg, Mn, and Pb levels in karst soils threaten local environments by impairing soil functions, reducing vegetation health, and increasing risks to human populations via agricultural and hydrological pathways [8,29]. These characteristics make karst regions priority areas for the detailed assessment and management of heavy metal contamination.

Hyperaccumulators are promising phytoremediation tools for extracting heavy metals from contaminated soils, offering an eco-friendly alternative to conventional methods [12,30]. This process involves plants absorbing metals through their roots and translocating them to aboveground tissues, thereby effectively reducing soil concentrations over multiple harvests [31]. The mechanism of hyperaccumulation typically includes enhanced root uptake via specific transporters, such as Natural Resistance-Associated Macrophage Protein and Zinc-Regulated Transporter family proteins, followed by chelation with ligands such as nicotianamine or phytochelatins to detoxify metals intracellularly [32,33]. Metals are then sequestered in vacuoles or bound to cell walls, preventing toxicity while enabling high accumulation [34]. Translocation to shoots is facilitated by xylem loading, which is often promoted by symbiotic microbes or root exudates that mobilize soil metals [14,35]. Typical hyperaccumulator species, such as Noccaea caerulescens, Pteris vittata, and Arabidopsis halleri, thrive in metalliferous soils, especially in Karst ecosystems [13,36,37]. In contrast, ordinary plants limit heavy metal accumulation to avoid stress, such as As below 5 mg kg dry weight and Cd below 3 mg kg⁻¹ [38]. However, hyperaccumulators can exceed 100 mg kg⁻¹ for Pb and Hg without growth inhibition, showcasing exceptional tolerance [38,39]. Therefore, hyperaccumulators can be considered as mediators of heavy metal-contaminated soils to facilitate the reuse of tailing stands in newly built parks.

Rapid urbanization and industrial legacy in many Chinese cities have left numerous underutilized or contaminated sites, including former tailing areas with the potential for ecological restoration. In Tongren, Guizhou Province, an abandoned tailing stand presents an opportunity for innovative urban greening through its conversion into a public park. However, the persistence of potentially toxic elements associated with past land use remains a critical concern for safe land redevelopment. The novelty of this study lies in its integrated evaluation of reclaimed mercury tailings land as a functional urban park within a karst ecosystem, combining heavy metal risk assessment with vegetation community dynamics under long-term phytoremediation conditions. The overall objective of this study was to demonstrate the feasibility of reusing tailing stands as functional urban parks by assessing heavy metal contamination risks. Specifically, we aim to: (1) quantify the concentrations of As, Cd, Hg, Mn, and Pb in soils and dominant plant species to determine whether contamination persists at levels posing ecological or human health risks, and (2) characterize soil–plant interactions and trace element transfer patterns within the developing botanic communities, in comparison with reference ordinary parks in Guiyang. These findings provide scientific evidence to guide safe phytoremediation and park development strategies in similar post-industrial urban sites.

2. Materials and Methods

2.1. Study Sites

This study was conducted as a field investigation of Hg tailings in Wanshan (27.50° N, 109.18° E), Guizhou, Northwest China. Land over these tailings has been converted to a park for recreational use by planting hyperaccumulating plant species for years. This location was taken as the objective with comparisons of another 30 parks distributed along an east–west geographical gradient (Figure 1). This transect was placed linking the tailings region (eastern end) to the western region, which was highly urbanized in the Huaxi District, Guiyang. Therefore, the 31 places were distributed from a gradient from a highly urbanized region (west) to a remote rural area (east); hence, parks in the western part had a higher dose of nature and higher quality of visits with a longer chronology than those in the east around the tailings.

The studied soils were shallow karst-derived soils with heterogeneous physical structures and variable rock fragment contents typical of carbonate landscapes in southwest China. Physically, the soils exhibited thin profiles, patchy soil–rock distribution, moderate compaction, and relatively limited water-holding capacity. Chemically, the soil pH ranged from slightly acidic to weakly alkaline, generally between 6.4 and 7.5.

2.2. Field Investigation

Ten plots were randomly set at each location, and each plot had an area of 600 m² (20 m × 30 m). Each plot was divided into six 100-m² subplots (10 m × 10 m), which were used as the basic sampling units for collecting soil and plant samples and investigating vegetative community structures. Plant species were selected based on three considerations: (1) dominance and frequency of occurrence within the investigated parks, (2) ecological adaptability to karst and mining-affected environments, and (3) previously reported tolerance or hyperaccumulation potential for the chosen heavy metals. Vegetative communities were analyzed for their structure in terms of coverage ratio, above-ground height, diameter at basal stand, species names, and numbers of individuals or clusters. Plant diversity was calculated using two variables: the Simpson index (D) and the Shannon–Wiener index (H′) [40,41]:

$$D=1-{\sum }_{i=1}^j{\displaystyle \frac{n\left(n-1\right)}{N\left(N-1\right)}}$$

(1)

$$H^{\prime }=-{\sum }_{i=1}^j{p}_i\mathrm{l}\mathrm{n}\left(p_i\right)$$

(2)

where n is the number of objective plant species, N is the total number from the start of i to the final j, and p_i is the proportion of individuals belonging to the i-th species to the total of j. The plant species are summarized in Appendix A with their common English names and Latin names.

Two soil cores were randomly sampled from a subplot using an earth drill (10 cm inner diameter) to a depth of 5 cm belowground. The extracted soils were initially screened to remove animal residuals, stones, and plant roots, and mixed to form a bulk sample per plot. Aboveground parts of plants were sampled per species per plot, which were bulked into two samples per subplot covering all essential species.

Vegetative plants were defined by their accumulation ability. Super-accumulators are plant species capable of absorbing and tolerating exceptionally high concentrations of heavy metals in their aboveground tissues without severe growth inhibition. These species possess specialized physiological mechanisms for metal uptake, translocation, sequestration, and detoxification, and are therefore considered valuable for the phytoremediation of contaminated soils. Non-remediators are plant species that exhibit relatively low uptake and accumulation of heavy metals and do not reach recognized hyperaccumulation thresholds in plant tissues. Such species generally lack strong metal extraction capacity and mainly reflect ordinary ecological adaptation rather than active phytoremediation potential. Appendix A

Table A1. Major plant species found in plots of this study.

English Common Name	Latin Name
Sweet wormwood	Artemisia annua
Chinese mugwort	Artemisia argyi
White clover	Trifolium repens
Mock strawberry	Duchesnea indica (syn. Potentilla indica)
Curly dock	Rumex crispus
Asian plantain	Plantago asiatica
Canadian horseweed	Erigeron canadensis
Annual fleabane	Erigeron annuus
Goosegrass	Eleusine indica
Beefsteak plant	Perilla frutescens
Oxeye daisy	Leucanthemum vulgare
Chinese silver grass	Miscanthus sinensis
Tatarian aster	Aster tataricus
Climbing groundsel	Senecio scandens
Multiflora rose	Rosa multiflora
Common vervain	Verbena officinalis
Asian cudweed	Gnaphalium affine
Wild radish	Raphanus sativus var. raphanistroides
Velvetleaf	Abutilon theophrasti
Siberian cocklebur	Xanthium sibiricum
Green foxtail	Setaria viridis
Indian pokeweed	Phytolacca acinosa
Oregano	Origanum vulgare
Cylindrical snake plant	Sansevieria cylindrica
Oriental false hawksbeard	Youngia japonica
Purple nutsedge	Cyperus rotundus
Knotweed/smartweed	Polygonum spp.
Japanese hedge-nettle	Stachys japonica
Amethyst-leaved isodon	Isodon amethystoides
Capillary wormwood	Artemisia capillaris
Water pepper	Polygonum hydropiper

shows English common names and their Latin names for major plant species found in plots of this study.

2.3. GPS-Based Spatial Mapping and Geostatistical Design

The geographic coordinates of all sampling plots were recorded in the field using a handheld Global Positioning System (GPS) receiver with sub-meter accuracy. The spatial locations of the 31 parks were mapped to characterize the east–west environmental gradient extending from the Wanshan Hg tailings region toward the urbanized parks of Guiyang. Sampling plots within each park were distributed to maximize spatial coverage and reduce the local clustering effects. The chosen plots were selected with standards that were evenly distributed in the transaction from Wanshan to Guiyang with similar stand attributes.

Geographic data were imported into ArcGIS 10.8 (Esri, Redlands, CA, USA) for spatial visualization and interpolation analysis. The spatial distribution patterns of soil heavy metals and vegetation attributes were evaluated using geostatistical approaches, including inverse distance weighting (IDW) interpolation and spatial autocorrelation analysis. Longitude and latitude variables were incorporated into the correlation analyses to assess geographical trends in contamination and vegetation responses.

The geostatistical framework enabled the identification of potential contamination hotspots, spatial heterogeneity, and ecological gradients associated with historical mining activities and urbanization intensity across the study region.

2.4. Chemical Analysis of Soil and Plant Samples

Soil and plant samples were transported to the laboratory on ice (0–4 °C) within eight hours. Soil samples were screened to pass a 2-mm sieve, dried at indoor temperature to a constant weight, ground, and passed through a 0.5-mm sieve. Plant samples were washed with distilled water, dried in an oven at 72 °C for 48 h and smashed by milling to pass through a 0.5 mm sieve.

The heavy metal content (As, Cd, Hg, Mn, and Pb) in soil and plant samples was determined using inductively coupled plasma–mass spectrometry (ICP-MS). For digestion, 0.50 g of dried soil or 0.20 g of dried plant material was accurately weighed and placed in Teflon digestion vessels. Samples were treated with a mixture of concentrated HNO₃ and H₂O₂ (analytical grade) and digested using a microwave digestion system (ETHOS UP, Milestone Srl, Sorisole, Italy). After digestion, the solutions were cooled, filtered if necessary, and diluted to a fixed volume using ultrapure water. The elemental concentrations were measured using an ICP-MS instrument (Agilent 7900 ICP-MS, Agilent Technologies, Tokyo, Japan). Calibration curves were prepared using multi-element standard solutions, and quality control included reagent blanks, duplicate samples, and certified reference materials to ensure analytical accuracy. The specific certified reference materials (CRMs) used for soil were GBW07405 (GSS-5, Chinese standard soil), and those for plants were GBW10015 (leaves in Bidens pilosa L.). The recovery rate (RR) was calculated as follows:

$$\mathrm{R}\mathrm{R}={\displaystyle \frac{\mathrm{M}\mathrm{V}}{\mathrm{C}\mathrm{V}}}\times 100\%$$

(3)

where MV and CV are the measured and certified values, respectively. The precision was estimated using the relative standard deviatio.

Mercury was analyzed in the collision/reaction mode to minimize spectral interference. The instrument detection limits for all elements were in the µg L⁻¹ range. The final heavy metal concentrations were calculated based on dilution factors and sample dry weights and expressed as mg kg⁻¹ dry weight for both soil and plant samples. The typical contents in soils and plants and their categories of standards are shown in

Table 1. Categories of heavy metal element contents according to standards in soils and plants.

Source	Element	pH Value, Standard (mg kg−1)	Reference
Soil	As	pH ≤ 5.5, 40; 5.5 < pH ≤ 6.5, 40; 6.5 < pH ≤ 7.5, 30; pH > 7.5, 25	GB 15618-2018 [42]
	Cd	pH ≤ 5.5, 0.3; 5.5 < pH ≤ 0.3, 40; 6.5 < pH ≤ 7.5, 0.3; pH > 7.5, 0.6
	Hg	pH ≤ 5.5, 1.3; 5.5 < pH ≤ 0.3, 1.8; 6.5 < pH ≤ 7.5, 2.4; pH > 7.5, 3.4
	Pb	pH ≤ 5.5, 70; 5.5 < pH ≤ 0.3, 90; 6.5 < pH ≤ 7.5, 120; pH > 7.5, 170
	Mn	Background: 657 mg kg−1, Chongqing; 583 mg kg−1, China; moderate standard: 170–1200 mg kg−1	Zhu et al. [43]
Plants	As	Normal range: 0.01–5.00 mg kg−1; Hyperaccumulator threshold, 100.00 mg kg−1	Ouyang et al. [38]
	Cd	Normal range: 0.20–3.00 mg kg−1; Hyperaccumulator threshold, 100.00 mg kg−1
	Pb	Normal range: 0.10–41.70 mg kg−1; Hyperaccumulator threshold, 1000.00 mg kg−1
	Hg	Normal range: 10 mg kg−1; Hyperaccumulator threshold, 100.00 mg kg−1	Qian [39]
	Mn	Radish, 136.10 mg kg−1; peppers, 94.37–1301.70 mg kg−1; Desmodium sequax, 533.9 mg kg−1; Paraixeris denticulata, 842.6 mg kg−1	Zhu et al. [43]

.

Local species that had been examined as hyperaccumulators included Ilex crenata Thunb., Arundinella hirta (Thunb.) Tanaka., Eremochloa ciliaris (L.) Merr., Imperata cylindrica (L.) P. Beauv., Saccharum arundinaceum (Retz.) Welker, Voronts. & E.A.Kellogg, Equisetum hyemale L., Miscanthus floridulus (Labill.) Warb. ex K.Schum. & Lauterb., Bidens pilosa L. 1753 [38], Pteris vittata L., Arthraxon hispidus (Thunb.) Makino., Fallopia multiflora (Thunb.) Haraldson., Equisetum ramosissimum Desf., Rumex acetosa L., Sedum emarginatum Migo [39], Polygonum lapathifolium Linn., Achyranthes bidentata Blume., Phytolacca americana Linn., Elsholtzia argyi Levl., Nephrolepis cordifolia (L.) Presl, Paraixeris denticulata (Houtt.) Nakai and Desmodium sequax Wall. [43].

2.5. Statistical Analysis

Data were tested for normal distribution and homogeneity of variance, and no abnormalities were indicated. Data was bulked to subplots and further averaged for plots, which led to ten replicated plots for each park. Analysis of variance (ANOVA) was used to detect differences in soil and plant parameters among the 31 plots. When significant effects were detected, multiple comparisons among means were performed using Tukey’s test, and statistically significant differences (p < 0.05) were indicated by different letters. The average means were calculated across plots per place, which were compared among places. The geographical gradient was analyzed between the longitude of the plots and their means. Correlation was also analyzed between pairs of variables among soil, heavy metals and vegetative community structures.

3. Results

3.1. Comparisons of Heavy Metal Contents in Soils

The ANOVA results in

Table 2. Comparisons of heavy metal contents (As, Cd, Hg, Mn, Pb) in soils distributed in the 31 places with means and analysis of variance (ANOVA) results.

Place	As	Cd	Hg	Mn	Pb
Mean ± SE (mg kg−1)
1	22.73 ± 0.22 B 1	0.27 ± 0.00 B	1.95 ± 0.03 A	865.21 ± 61.21 A	109.09 ± 2.40 A
2	14.16 ± 0.40 GFH	0.19 ± 0.01 FGDE	1.02 ± 0.03 DFE	577.60 ± 40.39 FEDC	48.22 ± 4.97 GHF
3	12.37 ± 0.42 JIH	0.17 ± 0.01 FGHI	0.81 ± 0.04 HGFE	444.59 ± 32.26 HFEDGI	43.99 ± 4.35 IGHF
4	11.10 ± 0.33 JLMK	0.14 ± 0.00 JKI	0.60 ± 0.06 HJI	372.92 ± 25.39 HLKGI	35.00 ± 4.57 KIGHMJL
5	11.20 ± 0.39 JLMK	0.14 ± 0.01 JKI	0.72 ± 0.03 HGI	435.72 ± 26.13 HFEJGI	36.10 ± 4.00 KIGHJL
6	7.17 ± 0.30 PO	0.09 ± 0.00 MN	0.28 ± 0.05 KML	269.26 ± 26.63 LKNJMI	18.83 ± 3.43 KMJL
7	6.68 ± 0.17 P	0.06 ± 0.00 N	0.16 ± 0.04 MN	271.18 ± 19.10 LKNJMI	22.15 ± 3.57 KIMJL
8	6.86 ± 0.24 P	0.06 ± 0.00 N	0.19 ± 0.05 ML	235.33 ± 22.86 LKNM	13.51 ± 3.28 M
9	10.35 ± 0.20 NLMK	0.13 ± 0.00 JK	0.54 ± 0.04 KJI	349.20 ± 27.84 HLKJMGI	34.11 ± 3.64 KIGHMJL
10	12.41 ± 0.22 JIH	0.15 ± 0.01 JKHI	0.79 ± 0.04 HGFI	453.25 ± 24.94 HFEDG	38.53 ± 2.63 KIGHJ
11	15.10 ± 0.19 EF	0.20 ± 0.00 FDE	1.07 ± 0.06 DCE	464.92 ± 24.73 HFEDG	64.40 ± 5.63 EDF
12	19.97 ± 0.23 C	0.22 ± 0.01 CD	1.66 ± 0.04 B	770.32 ± 46.37 BA	92.00 ± 4.99 BAC
13	13.13 ± 0.21 GIH	0.18 ± 0.01 FGHE	0.97 ± 0.06 DGFE	468.78 ± 20.41 HFEDG	49.65 ± 4.24 EGHF
14	14.68 ± 0.18 GF	0.19 ± 0.01 FGDE	1.07 ± 0.05 DCE	485.10 ± 19.78 FEDG	56.78 ± 4.69 EGDF
15	12.03 ± 0.23 JLIK	0.16 ± 0.01 JGHI	0.76 ± 0.05 HGFI	403.71 ± 12.09 HFKJGI	46.95 ± 4.11 GHF
16	12.12 ± 0.22 JIK	0.14 ± 0.00 JK	0.79 ± 0.04 HGFI	476.02 ± 10.04 FEDG	47.16 ± 4.16 GHF
17	9.98 ± 0.21 NM	0.12 ± 0.01 MKL	0.60 ± 0.04 HJI	290.91 ± 19.43 HLKNJMI	31.89 ± 4.04 KIGHMJL
18	10.24 ± 0.19 NLM	0.13 ± 0.00 KL	0.56 ± 0.04 HJI	331.19 ± 20.13 HLKNJMGI	40.29 ± 5.30 IGHJ
19	8.97 ± 0.16 NO	0.13 ± 0.01 KL	0.45 ± 0.04 KJL	320.12 ± 19.77 HLKNJMGI	23.18 ± 3.66 KIGHMJL
20	7.19 ± 0.15 PO	0.09 ± 0.01 MLN	0.35 ± 0.06 KMJL	261.43 ± 14.85 LKNJM	21.99 ± 4.21 KIGHMJL
21	5.93 ± 0.16 P	0.07 ± 0.01 N	0.14 ± 0.05 M	165.43 ± 12.50 N	22.60 ± 5.48 KIGHMJL
22	6.25 ± 0.18 P	0.08 ± 0.01 N	0.19 ± 0.06 ML	207.27 ± 11.41 LNM	16.55 ± 4.57 KML
23	7.01 ± 0.12 P	0.07 ± 0.01 N	0.17 ± 0.05 M	185.46 ± 12.35 NM	15.13 ± 3.86 ML
24	22.80 ± 0.15 B	0.27 ± 0.01 B	1.94 ± 0.03 A	810.99 ± 37.71 A	100.13 ± 3.10 BA
25	23.60 ± 0.14 B	0.27 ± 0.01 B	1.98 ± 0.05 A	826.33 ± 51.20 A	95.11 ± 4.61 BA
26	20.07 ± 0.18 C	0.24 ± 0.01 CB	1.74 ± 0.06 BA	700.38 ± 10.95 BAC	90.83 ± 2.95 BAC
27	17.97 ± 0.12 D	0.21 ± 0.01 CDE	1.32 ± 0.06 C	617.99 ± 11.87 BDC	77.74 ± 5.82 BDC
28	16.64 ± 0.10 ED	0.20 ± 0.01 FDE	1.22 ± 0.05 DC	580.99 ± 15.87 FEDC	70.90 ± 3.35 EDC
29	22.72 ± 0.14 B	0.24 ± 0.01 CB	1.79 ± 0.03 BA	806.59 ± 42.96 A	106.83 ± 2.61 A
30	22.00 ± 0.13 B	0.26 ± 0.01 B	1.91 ± 0.05 BA	818.02 ± 67.63 A	104.05 ± 2.91 A
31	29.22 ± 1.50 A	0.57 ± 0.02 A	1.88 ± 0.10 BA	611.18 ± 86.84 BEDC	100.76 ± 4.88 A
ANOVA
F value	326.25	247.04	163.5	40.43	57.9
p value	<0.0001	<0.0001	<0.0001	<0.0001	<0.0001

Note: ¹ Different capital letters (A–P) present significant difference according to Tukey’s test (α = 0.05).

indicate highly significant differences in heavy metal content (As, Cd, Hg, Mn, and Pb) across the 31 soil sampling locations. All F-values ranged from 40.43 (Mn) to 326.25 (As), and all p-values were lower than 0.0001, confirming substantial spatial variability. General trends among places showed elevated concentrations in places 1, 12, and 24–31, which are likely pollution hotspots, while places 6–8 and 20–23 exhibited the lowest levels, possibly indicating less impacted areas. Overall, As, Cd, and Pb contents followed similar patterns with peaks in industrialized or urban sites, whereas Hg and Mn varied more erratically. Specifically, for As, place 31 (29.22 mg kg⁻¹) exceeded all others; places 1, 24–25, 29–30 surpassed places 12, 26, and lower groups. For Cd, place 31 (0.57 mg kg⁻¹) was the highest, outperforming places 1, 24–25, 30. Hg was highest in places 1, 24–25 (1.95–1.98 mg kg⁻¹), above place 12 (1.66 mg kg⁻¹). Mn peaks in places 1, 24–25, and 29–30 (807–865 mg kg⁻¹) were higher than those in place 12 (770 mg kg⁻¹). Pb was elevated in places 1, 29–31 (101–109 mg kg⁻¹), exceeding that in place 12 (92 mg kg⁻¹).

As shown in Figure 2, the soil pH values across the 31 sampled locations showed significant variation. The highest pH value, indicating the most alkaline soil, was recorded at location 31, with a value of 7.53. Other locations with notably high pH values included locations 24, 25, 29, and 30. In contrast, the most acidic soil conditions were found at sites 21 and 23, which had the lowest pH values of approximately 6.46. The general trend across the locations revealed lower pH values in the middle-numbered places, with higher values observed at both the beginning and the end of the sampled range.

3.2. Comparisons of Heavy Metal Contents in Plants

The ANOVA results in

Table 3. Comparisons of heavy metal content (As, Cd, Hg, Mn, Pb) in plants distributed in the 31 places with means and analysis of variance (ANOVA) results.

Place	As		Cd		Hg		Mn		Pb
	Mean ± SE (mg kg−1)
1	14.05 ± 1.73	A 1	2.48 ± 0.06	A	12.51 ± 0.50	BDEC	254.44 ± 2.35	A	32.93 ± 0.79	BAC
2	7.70 ± 1.12	EBDGCF	1.44 ± 0.05	FGH	9.50 ± 0.44	HIG	162.73 ± 1.83	B	20.82 ± 0.67	GHI
3	4.87 ± 0.73	EDGHCF	1.22 ± 0.04	IJGH	8.75 ± 0.38	HJKIG	135.21 ± 1.74	GF	18.01 ± 0.63	JHI
4	3.87 ± 1.00	EGHF	1.09 ± 0.04	JKL	8.55 ± 0.42	HJKI	113.79 ± 1.43	IHJ	16.95 ± 0.61	JKI
5	5.84 ± 1.45	EDGHCF	1.12 ± 0.04	IJKL	8.76 ± 0.42	HJKIG	123.53 ± 1.92	L	17.17 ± 0.62	JKHI
6	2.19 ± 0.45	GH	0.55 ± 0.04	O	6.25 ± 0.41	MNL	85.34 ± 1.40	IKLJ	10.98 ± 0.63	NM
7	0.93 ± 0.19	H	0.46 ± 0.04	O	5.41 ± 0.34	MN	75.39 ± 1.48	M	9.06 ± 0.63	N
8	2.45 ± 0.44	GH	0.48 ± 0.04	O	5.38 ± 0.33	MN	77.09 ± 1.26	M	9.38 ± 0.63	N
9	3.36 ± 0.57	GHF	0.98 ± 0.04	MJKL	7.98 ± 0.36	JKIL	115.35 ± 1.72	M	15.98 ± 0.63	JKL
10	6.42 ± 1.51	EBDGHCF	1.15 ± 0.04	IJK	9.28 ± 0.38	HJIG	133.13 ± 2.30	KL	17.98 ± 0.63	JKHI
11	4.35 ± 0.66	EDGHF	1.48 ± 0.04	FG	10.63 ± 0.43	HFEG	156.32 ± 1.92	IKHJ	23.13 ± 0.77	GF
12	12.63 ± 1.64	BA	2.15 ± 0.04	BC	12.77 ± 0.44	BDEC	229.14 ± 2.64	GF	32.23 ± 0.80	BC
13	6.66 ± 1.25	EBDGHCF	1.37 ± 0.04	IGH	9.68 ± 0.37	HFIG	184.08 ± 2.11	C	20.04 ± 0.68	GHI
14	5.76 ± 1.52	EDGHCF	1.44 ± 0.04	FGH	10.76 ± 0.42	HFDEG	184.58 ± 2.03	ED	21.02 ± 0.68	GH
15	3.60 ± 0.62	EGHF	1.18 ± 0.04	IJKH	9.50 ± 0.40	HIG	148.06 ± 2.01	ED	17.68 ± 0.67	JKHI
16	2.34 ± 0.23	GH	1.14 ± 0.04	IJKL	9.39 ± 0.41	HJIG	136.93 ± 2.19	GH	17.35 ± 0.67	JKHI
17	3.45 ± 0.99	EGHF	0.88 ± 0.04	MNL	7.18 ± 0.36	MJKNL	110.23 ± 1.54	IH	14.63 ± 0.67	JKLM
18	3.68 ± 0.84	EGHF	0.95 ± 0.04	MKL	8.68 ± 0.41	HJKIG	117.95 ± 1.66	L	15.30 ± 0.67	JKL
19	1.82 ± 0.32	GH	0.82 ± 0.04	MN	7.68 ± 0.37	JKIL	106.98 ± 1.92	KLJ	13.97 ± 0.67	KLM
20	2.23 ± 0.43	GH	0.68 ± 0.04	ON	7.00 ± 0.37	MKNL	85.74 ± 1.62	L	12.30 ± 0.67	NLM
21	1.99 ± 0.45	GH	0.44 ± 0.04	O	5.23 ± 0.32	N	79.22 ± 1.38	M	8.70 ± 0.75	N
22	1.18 ± 0.35	H	0.50 ± 0.04	O	5.13 ± 0.30	N	78.43 ± 1.49	M	9.37 ± 0.75	N
23	1.07 ± 0.13	H	0.45 ± 0.04	O	4.94 ± 0.28	N	78.93 ± 1.31	M	8.81 ± 0.75	N
24	12.37 ± 1.51	BA	2.52 ± 0.04	A	14.15 ± 0.48	BA	256.37 ± 2.09	M	35.56 ± 0.65	BA
25	10.58 ± 1.31	BAC	2.55 ± 0.04	A	15.08 ± 0.47	A	275.78 ± 1.81	B	36.41 ± 0.59	A
26	12.06 ± 1.50	BA	2.11 ± 0.04	DC	12.92 ± 0.42	BDAC	216.20 ± 1.81	C	31.20 ± 0.81	DC
27	10.49 ± 2.12	BDAC	1.86 ± 0.04	DE	11.85 ± 0.41	FDEC	195.37 ± 1.80	D	28.06 ± 0.81	ED
28	7.50 ± 1.27	EBDGCF	1.67 ± 0.04	FE	10.94 ± 0.42	FDEG	167.71 ± 2.02	EF	25.32 ± 0.82	EF
29	9.64 ± 1.42	EBDAC	2.38 ± 0.04	BA	14.05 ± 0.47	BAC	255.53 ± 1.77	B	35.41 ± 0.65	BA
30	12.11 ± 1.71	BA	2.41 ± 0.04	BA	14.38 ± 0.47	BA	250.88 ± 1.98	B	35.73 ± 0.64	BA
31	9.39 ± 1.96	EBDACF	0.66 ± 0.14	ON	7.52 ± 0.81	MJKIL	169.37 ± 16.26	EF	9.46 ± 1.81	N
ANOVA
F value	12.11		191.46		49.7		331.52		148.83
p value	<0.0001		<0.0001		<0.0001		<0.0001		<0.0001

Note: ¹ Different capital letters (A–O) present significant difference according to Tukey’s test (α = 0.05).

demonstrate highly significant differences across all 31 locations for the five heavy metals, with p-values consistently falling below 0.0001. Manganese showed the highest overall concentrations, often exceeding 100 mg kg⁻¹, whereas Cd levels remained the lowest across the study areas. General trends indicated that Place 25 frequently exhibited the highest contamination, particularly for Mn (275.78 ± 1.81 mg kg⁻¹), Hg (15.08 ± 0.47 mg kg⁻¹), and Pb (36.41 ± 0.59 mg kg⁻¹). Specifically, Place 1 showed a higher As content (14.05 ± 1.73 mg kg⁻¹) than Place 24 (12.37 ± 1.51 mg kg⁻¹), and Place 25 contained more Cd than Place 12. Conversely, Places 7, 21, and 23 consistently showed the lowest levels of nearly all heavy metals.

Table 4. Categories of heavy metal content (As, Cd, Hg, Mn, and Pb) in plants subjected to the 31 places.

Places	As	Cd	Hg	Mn	Pb
1	[S] 1	[N] 1	[S]	[N]	[N]
2	[S]	[N]	[N]	[S]	[N]
3	[N]	[N]	[N]	[S]	[N]
4	[N]	[N]	[N]	[S]	[N]
5	[S]	[N]	[N]	[S]	[N]
6	[N]	[N]	[N]	[S]	[N]
7	[N]	[N]	[N]	[S]	[N]
8	[N]	[N]	[N]	[S]	[N]
9	[N]	[N]	[N]	[S]	[N]
10	[S]	[N]	[N]	[S]	[N]
11	[N]	[N]	[S]	[S]	[N]
12	[S]	[N]	[S]	[N]	[N]
13	[S]	[N]	[N]	[N]	[N]
14	[S]	[N]	[S]	[N]	[N]
15	[N]	[N]	[N]	[S]	[N]
16	[N]	[N]	[N]	[S]	[N]
17	[N]	[N]	[N]	[S]	[N]
18	[N]	[N]	[N]	[S]	[N]
19	[N]	[N]	[N]	[S]	[N]
20	[N]	[N]	[N]	[S]	[N]
21	[N]	[N]	[N]	[S]	[N]
22	[N]	[N]	[N]	[S]	[N]
23	[N]	[N]	[N]	[S]	[N]
24	[S]	[N]	[S]	[N]	[N]
25	[S]	[N]	[S]	[N]	[N]
26	[S]	[N]	[S]	[N]	[N]
27	[S]	[N]	[S]	[N]	[N]
28	[S]	[N]	[S]	[S]	[N]
29	[S]	[N]	[S]	[N]	[N]
30	[S]	[N]	[S]	[N]	[N]
31	[S]	[N]	[N]	[S]	[N]

¹ [S], super-accumulator level; [N], non-remediator level. The standard is shown in Table 1.

reveals that while most sites are categorized at “non-remediator levels” [N] for Cd and Pb, many locations, such as Places 24 through 30, reach “super-accumulator levels” [S] for As and Hg. This indicates a trend in which certain zones act as significant sinks for specific toxic elements.

3.3. Vegetative Community Structure

Table 5. Comparisons of vegetative community attributes in the 31 locations with means and analysis of variance (ANOVA) results.

Place	Coverage (%)		Height (cm)		Diameter (mm)		Simpson Index		Shannon–Wiener Index
Mean ± SE
1	81.30 ± 1.13	A 1	14.10 ± 0.17	A	13.96 ± 0.11	A	0.95 ± 0.00	AB	3.01 ± 0.01	A
2	63.50 ± 0.73	B	7.70 ± 0.11	EBDGCF	11.20 ± 0.06	GFH	0.90 ± 0.00	HIG	2.68 ± 0.01	CB
3	61.30 ± 0.95	B	4.87 ± 0.07	EDGHCF	12.37 ± 0.04	JIH	0.88 ± 0.00	HJKIG	2.52 ± 0.01	C
4	43.10 ± 1.05	C	3.87 ± 0.10	EGHF	11.10 ± 0.03	JLMK	0.86 ± 0.00	HJKI	2.41 ± 0.01	CD
5	42.40 ± 0.95	C	5.84 ± 0.15	EDGHCF	11.20 ± 0.04	JLMK	0.88 ± 0.00	HJKIG	2.31 ± 0.01	CD
6	43.10 ± 1.40	C	2.19 ± 0.05	GH	7.17 ± 0.03	PO	0.63 ± 0.00	MNL	1.95 ± 0.01	FG
7	35.30 ± 1.48	D	0.93 ± 0.02	H	6.68 ± 0.02	P	0.54 ± 0.00	MN	1.68 ± 0.01	GH
8	37.00 ± 1.26	D	2.45 ± 0.04	GH	6.86 ± 0.02	P	0.54 ± 0.00	MN	1.83 ± 0.01	FG
9	35.30 ± 1.72	D	3.36 ± 0.06	GHF	10.35 ± 0.02	NLMK	0.80 ± 0.00	JKIL	2.21 ± 0.01	DE
10	33.10 ± 2.30	D	6.42 ± 0.15	EBDGHC	12.41 ± 0.02	JIH	0.93 ± 0.00	HJIG	2.45 ± 0.01	CD
11	36.30 ± 1.92	D	4.35 ± 0.07	EDGHF	15.10 ± 0.02	EF	1.06 ± 0.01	HFEG	2.65 ± 0.01	CB
12	22.91 ± 2.64	E	12.63 ± 0.16	BA	19.97 ± 0.02	C	1.28 ± 0.00	BDEC	2.85 ± 0.01	B
13	18.40 ± 2.11	F	6.66 ± 0.13	EBDGHC	13.13 ± 0.02	GIH	0.97 ± 0.01	HFIG	2.35 ± 0.01	CD
14	18.40 ± 2.03	F	5.76 ± 0.15	EDGHCF	14.68 ± 0.02	GF	1.08 ± 0.00	HFDEG	2.51 ± 0.01	C
15	14.80 ± 2.01	G	3.60 ± 0.06	EGHF	12.03 ± 0.02	JLIK	0.95 ± 0.00	HIG	2.33 ± 0.01	CD
16	13.60 ± 2.19	G	2.34 ± 0.02	GH	12.12 ± 0.02	JIK	0.94 ± 0.00	HJIG	2.22 ± 0.01	DE
17	11.00 ± 1.54	H	3.45 ± 0.10	EGHF	9.98 ± 0.02	NM	0.72 ± 0.00	MJKNL	1.96 ± 0.01	FG
18	11.70 ± 1.66	H	3.68 ± 0.08	EGHF	10.24 ± 0.02	NLM	0.87 ± 0.00	HJKIG	2.22 ± 0.01	DE
19	10.60 ± 1.92	H	1.82 ± 0.03	GH	8.97 ± 0.02	NO	0.77 ± 0.00	JKIL	1.98 ± 0.01	FG
20	8.57 ± 1.62	I	2.23 ± 0.04	GH	7.19 ± 0.01	PO	0.70 ± 0.01	MKNL	1.76 ± 0.01	GH
21	7.92 ± 1.38	I	1.99 ± 0.05	GH	5.93 ± 0.02	P	0.52 ± 0.01	N	1.48 ± 0.01	H
22	7.84 ± 1.49	I	1.18 ± 0.04	H	6.25 ± 0.02	P	0.51 ± 0.01	N	1.52 ± 0.01	H
23	7.89 ± 1.31	I	1.07 ± 0.01	H	7.01 ± 0.01	P	0.49 ± 0.00	N	1.41 ± 0.01	H
24	25.64 ± 2.09	E	12.37 ± 0.15	BA	22.80 ± 0.02	B	1.94 ± 0.00	BA	3.02 ± 0.01	A
25	27.58 ± 1.81	E	10.58 ± 0.13	BAC	23.60 ± 0.01	B	1.98 ± 0.01	A	3.05 ± 0.01	A
26	21.62 ± 1.81	E	12.06 ± 0.15	BA	20.07 ± 0.02	C	1.74 ± 0.01	BDAC	2.84 ± 0.01	B
27	19.54 ± 1.80	F	10.49 ± 0.21	BDAC	17.97 ± 0.01	D	1.32 ± 0.01	FDEC	2.58 ± 0.01	BC
28	16.77 ± 2.02	G	7.50 ± 0.13	EBDGCF	16.64 ± 0.01	ED	1.22 ± 0.01	FDEG	2.42 ± 0.01	CD
29	25.55 ± 1.77	E	9.64 ± 0.14	EBDAC	22.72 ± 0.01	B	1.79 ± 0.00	BAC	3.01 ± 0.01	A
30	25.09 ± 1.98	E	12.11 ± 0.17	BA	22.00 ± 0.01	B	1.91 ± 0.00	BA	3.02 ± 0.01	A
31	16.94 ± 16.26	G	9.39 ± 0.20	EBDACF	29.22 ± 0.15	A	1.88 ± 0.01	MJKIL	3.10 ± 0.01	A
ANOVA
F value	40.43		12.11		326.25		163.5		247.04
p value	<0.0001		<0.0001		<0.0001		<0.0001		<0.0001

Note: ¹ Different capital letters (A–P) present significant difference according to Tukey’s test (α = 0.05).

demonstrates pronounced differences in vegetative community attributes among the 31 places, supported by highly significant ANOVA results (p < 0.0001 for all variables). Vegetation coverage varied substantially, ranging from very low values below 10% in some places to more than 80% in others. Mean plant height showed a wide gradient, from less than 1 cm at the lowest sites to over 14 cm at the tallest sites, whereas stem diameter ranged from approximately 6 mm to more than 29 mm. Biodiversity also differed strongly among locations, with Simpson’s index values spanning from approximately 0.49 to nearly 2.0 and Shannon–Wiener index values ranging from approximately 1.4 to above 3.1. Together, these results indicate contrasting community complexity and evenness across sites.

3.4. Changes Along the Geographical Gradient

Figure 3 shows the correlations between longitude and soil properties, plant elemental content, vegetation structure, and diversity indices. Most heavy metal elements (As, Cd, Hg, Mn, Pb) in soils and plants displayed strong positive correlations with longitude (R = 0.84–1.00, p ≤ 10⁻⁹), indicating increasing concentrations eastward. Soil pH also increased significantly (R = 0.94). In contrast, vegetation height, coverage, diameter, Simpson, and Shannon indices showed strong negative correlations (R ≈ −0.92 to −0.97, p ≤ 10⁻¹³), reflecting declining structural complexity and diversity along the longitude gradient.

3.5. Correlations Among Variables

As shown in Figure 4, the correlation heatmap illustrates the strong relationships among the soil, plant, and community variables. Soil and plant heavy metals (As, Cd, Hg, Mn, Pb) and soil pH showed high positive intercorrelations (R ≈ 0.72–1.00), indicating closely coupled accumulation patterns. In contrast, these variables were strongly negatively correlated with community traits, including height, coverage, diameter, Simpson, and Shannon indices (R ≈ −0.65 to −0.97). Community structural and diversity metrics were highly positively correlated (R ≥ 0.95), suggesting coordinated responses within plant communities.

4. Discussion

4.1. Heavy Metal Contents in Soils

The upper bounds of As and Cd were comparable to or lower than those reported for post-mining parks and remediated tailings in subtropical regions, where concentrations often approach regulatory limits, even after management [44,45]. The Hg and Pb contents were notably lower than those documented for legacy Hg mining soils in Wanshan prior to remediation, where Hg frequently exceeded 3–5 mg kg⁻¹ and Pb surpassed 150 mg kg⁻¹ [46]. Manganese values were largely within the regional background to moderate ranges, aligning with karst soil surveys that emphasize strong lithological control rather than anthropogenic enrichment [47].

The results also demonstrated significant spatial variability among sites; however, the absolute values indicated a declining gradient of As, Cd, and Pb from historically impacted eastern locations toward urbanized western parks. This spatial pattern resembles the trends observed in rehabilitated mining corridors, where the distance from the source and long-term vegetation cover reduce soil metal loads [1,3]. The results further showed that sites hosting hyperaccumulating plant assemblages maintained lower or stabilized soil metal pools relative to ordinary parks, which is a pattern consistent with studies reporting sustained extraction and containment effects under field conditions [2,48]. Compared with similar phytoremediation trials in Asia and Africa, the observed As, Cd, and Hg levels were distinctly lower, supporting the effectiveness of prolonged plant-mediated removal [4,49].

The strong positive associations among soil As, Cd, Hg, Mn, and Pb, and their corresponding plant concentrations, suggest that residual metal availability still regulates plant uptake across the east–west gradient. This was observed despite the overall reduction in contamination levels after reclamation. The coupled responses between soil and plant variables further indicate that tolerant vegetation assemblages maintained active metal sequestration and stabilization functions under field conditions, which agrees with observations from mining tailings and contaminated urban soils, where hyperaccumulator communities exhibited coordinated soil–plant transfer patterns [1,45]. In contrast, the pronounced negative correlations between metal concentrations and vegetation diversity indicate that residual contamination continued to constrain community development at highly impacted sites, even when soil concentrations remained below regulatory thresholds.

Soil samples were collected from surface soils, which may contradict the consideration that heavy metals vertically migrate via fissures in karst soils. This can be obstructed by phytoremediation, as heavy metals can be absorbed by plants and returned to the surface soil through litter. It was reported that cities in the karst regions of southwestern China had higher heavy metals in surface soils than in deeper soils [50], including observations in Yinjang County [51] and Liuzhou City [52]. This demonstrates that the collection of surface soils is a reasonable method for representing the level of heavy metal contamination in soils to the greatest extent. This may further activate speculation that deeper soils may be a considerable candidate, or even a better choice, for soil sampling collection. This is argued to be an improper arrangement. First, the stands for soil collection were vegetative, and plants extracted heavy metals from deep soils with root proliferation in belowground layers. The absorbed heavy metals can be returned to topsoil through litter decomposition. Second, the investigated region was contaminated by tailings, which caused heavy metal input to the surface earlier than to the deeper soils. Hence, surface soils should exhibit higher levels of contamination, which are more representative than deeper soils. Finally, topsoil posed a higher risk of heavy metal pollution than deeper soils, as anthropogenic activities rarely affect deep soils [53]. Overall, the collection of topsoil was confirmed to be representative of heavy metal contamination in our study.

At the regional scale, the combined evidence confirmed that soil heavy metals did not exceed safety limits across the investigated parks. The consistently moderated concentrations, together with spatial patterns favoring remediated sites, endorsed the functional role of hyperaccumulation in reducing legacy contamination and maintaining soil quality suitable for long-term park use in the present study. Nevertheless, we recognize that total concentration alone cannot fully represent environmental or human health risk. To address this, this study integrated ecological indicators, including plant uptake, bioaccumulation behavior, and vegetation responses, as complementary lines of evidence. Future studies should incorporate bioavailability analyses and multi-pathway exposure models to provide a more comprehensive and mechanistic risk assessment framework.

From the perspective of environmental toxicology, total metal concentration alone does not accurately reflect toxicity or mobility; hence, sequential extraction methods are recommended. High-resolution in situ techniques, such as planar optodes and Diffusive Gradients in Thin Films (DGTF), are also recommended for dynamically assessing metal bioavailability and release fluxes at the soil–root interface [54]. However, it is important to clarify that the primary aim of this study was not to directly reflect the true toxicity or mobility of metals, but rather to evaluate the overall feasibility of reusing contaminated tailings land as urban parks through a comprehensive assessment of residual metal levels and vegetation responses. Specifically, this study focused on determining whether the total concentrations of these elements remain within regulatory safety thresholds and whether plant communities can be sustainably established under such conditions. Therefore, total metal concentration serves as a practical and standardized indicator for environmental risk screening and land-use suitability, rather than a proxy for bioavailability or dynamic behavior.

4.2. Heavy Metal Contents in Plants

Plant heavy metal content exhibited clear spatial differentiation along the longitudinal gradient, while remaining within reasonable physiological ranges. The concentrations of As, Cd, Hg, Mn, and Pb in the aboveground tissues increased gradually from western urban parks to the eastern sites adjacent to the former Hg mining area, mirroring the soil gradient. This spatial pattern agrees with observations from karst and post-mining landscapes, where plant metal loads closely track local geochemical backgrounds and legacy contamination intensities [25,55]. However, even at the eastern sites, the mean plant As and Cd contents were comparable to those reported for dominant vegetation in remediated gold and lead mining areas, rather than values typical of highly contaminated systems [3,15].

The differences among places further highlight the role of site history and plant assemblages. Parks closer to Wanshan showed higher Hg and Mn accumulation, whereas Pb and Cd were more prominent in several urbanized western parks, reflecting mixed industrial and traffic sources. Similar inter-site contrasts have been reported for wild and managed plants growing along urban–rural gradients in southern China, where species composition strongly mediates metal uptake [17,56]. Across all sites, the Mn content in plants mostly fell within the ranges reported for tolerant karst species, supporting the view that Mn enrichment reflects natural availability rather than toxic overload [19].

Importantly, the metal concentrations in the plants did not exceed the commonly accepted normal ranges or hyperaccumulation thresholds. These values were generally lower than those reported for known hyperaccumulators deliberately grown in metalliferous soils [55,57]. This pattern indicates that the screening and selection of planted species effectively limited excessive accumulation while maintaining the remediation capacity. Comparable strategies have been demonstrated to balance ecological safety and metal extraction in reclaimed parks and afforested mining lands [2,5]. Overall, the regional conclusion is that heavy metal content in plants remains within reasonable and manageable ranges, which confirms that appropriate plant screening can support safe park reuse in legacy mining landscapes.

4.3. Driving Effects of Heavy Metal on Vegetative Communities

The multivariate patterns indicate that heavy metal content in soils and plant tissues acted as important, but not exclusive, drivers of vegetative community structure across the parks studied. Soil As, Hg, and Pb displayed significant negative associations with species richness, canopy coverage, and diversity indices, whereas Cd and Mn exhibited weaker or non-significant relationships with these factors. These patterns suggest that highly toxic and persistent elements impose stronger ecological filtering than metals that are either essential at low doses or are more readily regulated by plants. Similar declines in diversity under elevated As, Hg, and Pb levels have been reported in mining-impacted parks and urban green spaces, where sensitive taxa are gradually excluded and tolerant species dominate the community assembly [45,58]. However, this study only calculated alpha diversity (i.e., Simpson and Shannon indices) and failed to robustly demonstrate the successional trajectories of vegetative communities under heavy metal stress. Beta diversity analyses should be considered in future studies, employing parameters such as non-metric multidimensional scaling (NMDS) or principal coordinates analysis (PCoA), which can show species turnover [59,60].

The correlative heatmap also revealed that community height, coverage, and diversity indices responded synchronously to changes in heavy metal concentrations, which supports the interpretation that vegetation recovery was linked to declining ecological stress along the geographical gradient. Similar inverse relationships between metal enrichment and plant structural performance have been reported in phytoremediation systems affected by mining activities, where increasing metal loads reduced vegetation complexity and biomass accumulation [44,45]. The highly positive correlations among the community indices further suggest that species establishment, canopy development, and diversity recovery progressed in a coordinated manner during the long-term reclamation.

The results further illustrate that plant metal contents track soil gradients only partially, indicating effective physiological regulation and species-specific tolerance. Essential micronutrients, such as Mn, showed a unimodal relationship with biomass and coverage. This reflects their role in photosynthesis and enzyme activation at moderate levels, while exerting stress only when the thresholds are exceeded [20],. In contrast, non-essential metals such As and Hg displayed consistently negative trends with structural attributes, which aligns with the reported disruptions of root growth and nutrient uptake in contaminated soils [61,62]. On the other hand, the east–west gradient was laid out to detect the difference in heavy metal contamination between post-mining lands and urbanized lands with parks built. This study was not designed to examine the contamination risk caused by residual heavy metals in the soil. This design was not used for testing urbanization responses, although plots were set along a rural-urban transect. Given that the tailings in Wanshan are planned to be rebuilt as a park in the future, the gradient can be taken as a comparison between the current scene and the future scenario.

The results highlight that the variation in community composition was explained jointly by heavy metals and site factors, with metal effects appearing secondary to habitat age and management intensity at several locations. The relatively weak correlations observed for Cd and Mn confirm that rare or low-level contamination did not necessarily translate into marked community degradation, particularly in the presence of tolerant or hyperaccumulating species. Comparable findings have been documented in reclaimed mine parks, where metal-tolerant assemblages maintain stable structures despite residual contamination [3,63].

Collectively, these results support the endorsement of reusing established vegetative communities for newly built parks. The limited and element-specific impacts of heavy metals, combined with plant tolerance and regulation capacity, indicate that existing communities can sustain ecological functions and landscape stability when contamination remains below critical thresholds and ongoing management is applied [4,64].

4.4. Limits of This Study

The current study had six limitations.

First, the methods rely on shallow surface soil sampling and measurements of aboveground plant tissues, which limits the representation of vertical metal heterogeneity and root–soil processes, particularly in karst systems with strong subsurface transport. Sampling was conducted during a single period, which restricts the evaluation of seasonal variation in metal uptake, plant growth, and community structure under changing environmental conditions.

Second, the results were mainly based on correlative analyses of soil metals, plant metals, and community attributes. Such relationships indicate consistent patterns but do not establish causality, especially where heavy metals covary with soil pH, park age, and management. The dominance of tolerant or planted species may further obscure subtle stress responses, implying that the rare or early ecological effects of low-level metal exposure could be underestimated.

Third, given that our current focus was to examine soil concentrations of heavy metals to evaluate the suitability of reusing local lands as parks, this study should have incorporated protocols that evaluate the risks of human exposure, such as the one proposed by the United States Environmental Protection Agency.

Fourth, we acknowledge that indices such as the Hakanson ecological risk index, bioconcentration factor (BCF), translocation factor (TF), pollution load index (PLI), and USEPA human health risk models could provide a more quantitative evaluation of ecological toxicity, metal transfer efficiency, and potential exposure risks. We suggest that subsequent investigations combine ecological indices with multipathway human exposure models and geostatistical analyses to improve the understanding of contamination dynamics, phytoremediation effectiveness, and long-term suitability for sustainable urban park development.

Fifth, studies on the mechanisms underlying heavy metal stabilization are still scarce. It is important to detect the contributions of microbial communities and enzyme activities in soils that reshape soil environments and community succession in future studies.

Finally, the overall design centered on a longitudinal gradient surrounding a former Hg mining area, which constrained extrapolation to other urban or post-industrial landscapes. Differences in historical management among reference parks introduce background variability, and the lack of long-term monitoring limits the assessment of delayed or cumulative ecological responses during continued park use.

5. Conclusions

This study applied a systematic field investigation combined with standardized chemical analyses to evaluate soil–plant interactions across a longitudinal gradient of parks surrounding a former Hg tailings site in Wanshan, China. Soil and dominant plant tissues were sampled using replicated plots, and heavy metal concentrations were quantified using ICP-MS to ensure consistency and comparability across sites. Statistical analyses identified spatial patterns and relationships among metals, vegetation structure, and geographical position, allowing an integrated assessment of remediation outcomes. The results demonstrated clear and consistent spatial trends in the distribution of heavy metals. Soil and plant concentrations of As, Cd, Hg, Mn, and Pb showed significant variability among parks and increased eastward toward historically impacted areas, whereas vegetation coverage, height, and diversity decreased along the same gradient. Despite these trends, the absolute concentrations remained moderate, and phytoremediation corresponded with stabilized soil conditions and structured communities. The soil heavy metal content remained below the established safety thresholds across all investigated parks. Plant tissues also exhibited metal concentrations within normal or controlled ranges, reflecting the deliberate selection of tolerant and hyperaccumulating species. These species effectively regulate uptake, reduce residual contamination, and prevent excessive accumulation in the above-ground biomass. Overall, the cleaned Hg tailings site in Wanshan almost met the ecological safety requirements for public green spaces; however, more conclusive conclusions should be made based on supporting data on deep soil profiles, groundwater monitoring, and long-term observational series. The combined reduction in soil metals and controlled plant accumulation supported the conclusion that the site was suitable for sustainable park use and long-term recreational development in the future.