Ten Years of Aided Phytostabilization in a Former Cu Mine: Evaluating Soil Recovery Under Different Amendment and Vegetation Strategies

1. Introduction

Mining and processing of mineral resources enable human socio-economic development. However, they also lead to significant socio-environmental degradation, particularly in open-pit operations and where environmental regulation, waste management, or post-closure restoration are inadequate [1]. Mining activities cause extensive environmental destruction, with severe and recurring impacts on soil, such as subsidence, sealing, generation of abrupt topography, compaction, erosion, and chemical contamination, particularly from potentially toxic elements (PTE) [1,2]. These elements, which include metals and metalloids such as As, Cd, Cr, Cu, Ni, Pb, Se and Zn (among others) occur naturally in the soil, but mining activity alters their concentration, distribution, and mobility, leading to potential environmental risk [3,4]. Often in metal mining, the exposure of sulfides promotes oxidation processes that may cause severe acidification in the environment, further increasing Fe and PTE solubility [5,6,7]. As a result, many metal mining soils exhibit high toxicity and infertility, with low organic matter content, low cation exchange capacity and water-holding capacity, and low biodiversity [8,9,10]. Global mining land use is estimated at 101,583 km² and, due to digital and energy transitions’ growing demand for strategic minerals, it is expanding at the faster rate ever in History [11]. Metal mining is a substantial and critical component of this footprint and represents one of the main pressures on global biodiversity decline [12].

The negative impacts of mining call for the implementation of sustainable restoration strategies. In this context, phytoremediation was proposed as an ecological restoration strategy aimed at reducing the environmental risk posed by degraded sites through the use of plants and related agronomic practices. It generally involves establishing an in situ vegetation cover that harnesses solar energy through photosynthesis. That vegetation can simultaneously reduce PTE availability and mobility, produce valuable biomass, and contribute to the development of soil functionality with minimal energy input [13,14,15]. Therefore, it represents a less intensive alternative – both economically and environmentally – than conventional techniques (e.g., excavation, containment, and chemical extraction of contaminants), and its implementation is supported by various scientific societies, such as the International Phytotechnology Society, the Society of Environmental Toxicology and Chemistry (SETAC), and the Society for Ecological Restoration [16]. However, its application requires site-specific studies, appropriate plant selection (species and intraspecific variation) and management techniques, as well as medium-to-long-term maintenance and monitoring [10,17,18].

Among phytoremediation strategies, phytostabilization aims to establish a stable vegetation cover to promote the retention of pollutants in the soil, thereby reducing their mobility, bioavailability, and dispersion into other environmental compartments [13,14]. Certain herbaceous species – for example, within the genus Agrostis – and woody species from genera such as Salix and Populus have drawn significant interest in phytostabilization projects due to their tolerance to PTE pollution, ability to thrive in oligotrophic soils, and high productivity [15,18,19]. Particular plant traits such as the restriction of pollutants’ translocation into above-ground biomass, or its retention in the rhizosphere or underground organs are specifically selected [10,17,20].

In degraded sites, adverse soil properties hinder the establishment of vegetation cover. Therefore, aided phytoremediation, that is phytoremediation assisted by soil amendments, has been proved as an effective strategy for improving soil conditions and mitigating the bioavailability of PTEs [13,14,21,22]. Liming agents, such as phosphates and apatite, Fe-, Al-, Mn-(hydr)oxides, biomass ashes, paper mill sludge, and organic wastes have been widely used in various remediation, including phytostabilization, trials [23,24,25,26]. These amendments can correct physical, chemical, and biological limitations through processes such as particle aggregation, moisture retention, acid neutralization, the addition of organic matter and nutrients, and/or the immobilization of PTEs (by adsorption, precipitation, organic complexation, or incorporation into poorly soluble mineral phases), among others. Among organic soil amendments, compost stands out for its multifunctionality: supplying carbon and nutrients, improving soil structure, buffering pH, and creating stable organometallic complexes [10,18,23,27]. In recent years, this approach has been expanded through the design of complex amendments and technosols derived from organic, mineral, or industrial byproducts and waste materials [28,29]. These amendments can be formulated to perform specific functions, such as regulating pH, supplying basic cations, increasing cation exchange capacity, improving water retention, or immobilizing PTEs [30]. Thus, the use of compost and designed technosols allows for soil restoration coupled with waste recovery opportunities, in line with the principles of the circular economy and carbon sequestration [31]. However, these materials may also contain PTEs or other hazardous components [32,33], so their effectiveness must be evaluated not only in terms of initial soil improvement but also with regard to the long-term stability of their effects.

The development of a vegetation cover also promotes gradual recovery of the edaphic ecosystem’s functionality. Plants physically stabilize the soil, reduce erosion, activate biogeochemical cycles, increase organic carbon inputs, and can improve soil structure and microbial activity [10,13,18,34]. When these processes are integrated, in an aided phytoremediation framework, with harvesting, the approach aligns with the concept of phytomanagement, understood as the sustainable management of contaminated sites that combines risk reduction, ecological restoration, and the generation of environmental or productive value [15,35,36]. In the case of woody species, their management in short-rotation systems can also facilitate the use of biomass for material or energy purposes [14,15], provided that the risk associated with the accumulation of PTEs is properly assessed.

On the other hand, time is an important factor to consider. Although amendments can lead to rapid improvements in pH, fertility, and PTE availability, the functional recovery of mine soils is typically slow and depends on complex environmental dynamics (e.g., plant-soil interaction, edaphic processes, weather patterns). Therefore, long-term monitoring is essential [15], especially in mine soils, which exhibit high spatial heterogeneity associated with both the original arrangement of materials and earthworks during mining operations and restoration efforts [9,18].

The effect of soil amendments and vegetation on soil properties in mining environments has been evaluated by various authors; however, most of these studies have been conducted under greenhouse conditions or over relatively short periods of time [10,13,24,28,37,38]. Long-term field studies are less common and, in many cases, are based on one-time sampling or chronosequences established across different areas [39,40]. In this context, periodic monitoring of the same plots is particularly useful for tracking processes, limiting the effect of spatial variability, and strengthening the robustness of the conclusions. This study evaluates the long-term evolution of a hyperacidic mine soil – poor in organic matter and nutrients but high in Cu – following the application of various aided phytostabilization strategies. Particularly, the study compares the effects of a compost-based organic amendment and two technosols on general soil properties, plant nutrients and available PTE fractions. It also analyzes the role of different vegetation covers in the system’s evolution. The initial hypothesis is that soil amendments immediately improve fertility conditions and reduce the availability of PTEs, but that the persistence of these effects depends on the type of amendment, the presence of vegetation, and the temporal evolution of the soil-plant system.

2. Materials and Methods

2.1. Location

The trial was conducted on a mine spoil heap in a former copper mine at Touro municipality (A Coruña, northwestern Spain; ETRS89 datum: 42° 52’ 34” N, 8° 20’ 40” W), being part of “PhytoSUDOE” network of phytomanaged sites (Phy2SUDOE Project, SOE4/P5/E1021). The climate is Atlantic oceanic, with average annual temperatures and precipitation of 13.0 °C and 1787 mm [41]. Geologically, the place is located in the Fornás unit, within Órdenes complex, in Galicia-Trás-os-Montes zone, consisting of diverse metamorphosed basalts (amphibolites) and siliciclastic sequences (schist rocks to paragneiss) attributed to the Precambrian–Paleozoic transition. The composition and grain size of the amphibolites vary greatly, but they generally contain amphibole (hornblende), chlorite, garnet (almandine), pyroxene, quartz, and biotite. Ore mineralization, consisting of pyrrhotite, chalcopyrite, sphalerite and pyrite, is mainly associated with coarse-grained, Ca-poor, Fe-rich garnet-bearing amphibolites, which gradually transition into unmineralized amphibolites and paragneiss [42]. Mining activity began in the early 1970s with the open-pit mining site of Arinteiro, later expanding to Vieiro, Bama, and Brandelos. Between 1973 and 1986, “Rio Tinto Patiño S.A.” extracted 21 Mt of ore materials (0.61 % de Cu), generating 38 Mt of excavation debris and 28 Mt of processing residues (differential flotation of chalcopyrite), which were deposited in mine spoil heaps and flotation tailings, respectively. The ore reserves have recently been estimated at 103 Mt (0.41% Cu) [43].

The oxidation of sulfides resulted in hyperacidic (pH 2 to 3) and hyperoxidizing (Eh > 500 mV) soils and waters, with high electrical conductivity (EC) and high concentration of sulfate, Fe and Al, and trace elements, which were assayed in different areas since the 1990s [44,45,46,47,48]. Specifically, this trial was conducted at the Arinteiro Site. The place showed signs of active oxidation near the surface, and plant growth was virtually nonexistent. The mine soil (Spolic Technosol (dystric, episkeletic)) [49] showed no horizon differentiation, had high gravel content, and was depleted in organic matter and nutrients; it was extremely acidic and had a low effective cation exchange capacity (CEC_e), dominated by Al³⁺ and H⁺ [49]. Furthermore, Cu content far exceeded the Generic Reference Level [50] of natural soils in the area.

2.2. Experimental Design

2.2.1. Application of Soil Amendments

The experiment began in spring 2011. The selected area of study (0.1 ha) was mechanically tilled (50 cm) and parceled into three plots, one amendment each: mature organic compost (PC) (made from urban sewage sludge and bark chips) and two technosol mixtures (PT1 and PT2). The technosols were produced by combining organic waste (anaerobic and aerobic municipal sewage sludge) and alkaline inorganic by-products (fly ash from biomass combustion and paper mill dregs). Each technosol contained a repurposed industrial by-product as a specific conditioner: Fe (hydr)oxides rich mixture (T1) and Al (hydr)oxides rich mixture (T2). All three soil amendments were manufactured, characterized, and incorporated by “Tratamientos Ecológicos del Noroeste (TEN S.L.)”, a local SME company dedicated to waste recycling and the production of organic fertilizers and soil amendments. The application rates were equivalent to 250 kg of amendment per Mg of minesoil in the first 30 cm. The plantation subplots were laid out, but the selected vegetation was not introduced immediately. Instead, the soil mixture was allowed to settle and equilibrate for approximately six weeks. The surrounding unamended mine tailings served as the negative amendment control (UNT).

2.2.2. Establishment of Plant Species

Each experimental plot was again parceled into 5 m × 5 m planting subplots, and three randomized replicates were established for two metal-tolerant crops and its control. On one hand, Agrostis capillaris (L.) cv. Highland was sown at a rate of 250 kg ha^-1 (AGR), using commercially available seeds. On the other hand, rooted clonal cuttings of Salix viminalis (L.) were planted at a rate equivalent to 10,000 cuttings ha^-1. These cuttings were obtained from local specimens naturally adapted to the environmental conditions (high levels of trace elements, climate, biological agents, etc.). In addition, three subplots were left unplanted (UNP) in each plot as negative planting control. The introduction of vegetation was not attempted on unamended soils, as plant development was severely impaired on these unamended soils. Between 2012 and 2015, established vegetation was harvested and all subplots mechanically weeded. The established plants were able to regenerate vegetatively after cutting, while the UNP areas were left to develop naturally thereafter.

2.3. Soil sampling and Analysis

Before applying any treatment, five soil samples (0 – 20 cm deep) were randomly collected from each plot; these were combined to obtain a composite sample and establish the initial soil conditions (t₀). To monitor subsequent soil changes, soil samples were similarly collected from each subplot (5 m × 5 m) and the surrounding UNT control in 2012, 2013, 2014, 2017, and 2021 (t₁, t₂, t₃, t₆, t₁₀, respectively). The samples were collected removing the litter layer and using unpainted stainless-steel shovels. They were stored and transported to the laboratory in polyethylene bags. Systematically, soil samples were dried at room temperature, sieved (< 2 mm), and homogenized for subsequent analysis, always using the equivalent analytical methods and equipment.

Soil pH was measured in H₂O ultrapure and 0.1 M KCl in suspension 1:2.5 (w/v) using a Metrohm 632 pH meter. Total C and N (C_T, N_T) were analyzed by combustion using an elemental analyzer (CHN-1000, LECO Corp., St. Joseph, MI). Exchangeable cations (Ca²⁺, Mg²⁺, Na⁺, K⁺, Al³⁺) were extracted with unbuffered 1 M NH₄Cl (1:20 w/v, 16 h equilibration) and measured by optical emission spectrometry (ICP-OES, Vista-PRO, Varian). The CEC_e was calculated as the sum of the exchangeable cations [51]. Additionally, NH₄Cl-extracted H⁺ was determined by titration whenever extraction pH was below 4.5. Available phosphorus (P_Olsen) was extracted using the Olsen [52] method (0.5 M NaHCO₃, pH 8.2, 1:20 w/v, 30 min agitation) and determined by colorimetry of the phosphomolybdate complex [53] using a Biochrom Libra S60 spectrophotometer. The pseudo-total elemental concentration ([E]_T) was determined in 0.5 g of ground soil, following oxidation (H₂O₂) and subsequent acid digestion (HNO₃:HCl, 1:3) in a microwave-accelerated reaction system (MarsXpress; CEM Corp., USA, or ETHOS, MILESTONE; Italy) in PFA Teflon vessels at 190 ºC, 9 bar, and 45 min. The readily available fraction ([E]_RA) was estimated by extraction with 0,1 M NaNO₃ or 1 M NH₄NO₃ (2:5 w/v soil:solution, 2 h agitation). PTE concentrations (Cd, Co, Cr, Cu, Mn, Ni, Pb, and Zn, both pseudo-total and readily available) were analyzed by ICP-OES (Vista-PRO, Varian). A certified soil standard (ISE 979 Renzina Soil) was used as quality control. Al and Fe (pseudo-total) were also analyzed after acid digestion.

2.4. Statistics

Initial data processing was performed using Microsoft Excel, and statistical analyses were conducted in RStudio (2026.04.0; R version 4.6.0). Unless otherwise specified, the threshold for statistical significance was set at p = 0.05.

The temporal dynamics of soil properties were analyzed using generalized linear models (GLM) based on generalized least squares (GLS), implemented with the package nlme [54,55] in R. In amended soils, the models included the factors Amendment and Plantation as categorical predictors, three levels both (Amendment: PC, PT1, and PT2; Plantation: UNP, AGR, and SAL), as well as Time as a continuous predictor. The effects of the categorical factors were estimated using zero-sum contrasts (contr.sum), while the temporal evolution was evaluated using polynomial contrasts (contr.poly). Since the inclusion of untreated soils (UNT) resulted in trivial differences that masked the effects between treatments, separate models were fitted for each response variable: one considering only the amended soils and another restricted to the UNT control.

All models were evaluated and adjusted iteratively to satisfy the assumptions of normality of the residuals, homoscedasticity, and minimization of Akaike’s Information Criterion (AIC). The diagnosis was performed through graphical and analytical scrutiny using the package performance [56]. Whenever necessary, time-dependent variance structures and/or logarithmic transformations of the response variables were incorporated. The final models were recalculated using restricted maximum likelihood (REML), and the overall explanatory fitness was estimated using a coefficient of determination (R²) calculated as one minus the ratio of residual to total observed variance.

The significance of the main effects and their interactions was assessed using marginal Wald χ² tests applied to the fitted GLS models, analogous to a Type III ANOVA. In addition to p-values, partial quadratic omegas (ω²) were calculated as approximate effect size estimators, i.e., the magnitude of the effect of each factor (and interaction) on the variance explained by the model. Marginal predictions and their confidence intervals (95%) were obtained using the package emmeans [57] and were plotted alongside the observed values.

Multivariate relationships among soil variables were explored using principal component analysis (PCA) in the FactoMineR environment [58,59,60]. The first three principal components were retained; together, they explained 74.0% of the total variance and were the only ones with an individual contribution greater than 10%. The variables and samples were plotted on two-dimensional plots defined by pairs of principal components. In addition, the distribution of the samples according to the factors Amendment, Plantation and Time was explored using labels, colors, and confidence ellipses.

3. Results

3.1. Short-Term Effects of Amendments on Soil Physicochemical Properties

The mine soil in the study area prior to amendment (PC t₀, PT1 t₀, and PT2 t₀, Table 1) was extremely acidic (pH_H2O 2.7 to 3.6, on average), low in organic matter (C_T ≤ 1.0%) and nutrients (N_T ≤ 0.10%, P_Olsen ≤ 4.0 mg kg^-1). The CEC_e was low, and the exchange complex was dominated by H⁺ and Al³⁺, with base saturation below 20% (on average). The pseudo-total contents of PTEs showed high spatial variability, but high concentrations of Mn_T (787 – 848 mg kg^-1) and Cu_T (567 – 630 mg kg^-1) were found, followed by Zn_T (117 – 132 mg kg^-1), Cr_T (89 – 103 mg kg^-1), Ni_T (40 – 51 mg kg^-1), Pb_T (16 – 18 mg kg^-1), and Cd_T (≤ 2 mg kg^-1). As for the readily available fractions, the highest values were observed for Cu (Cu_RA, 51 – 108 mg kg^-1) and Mn (Mn_RA, 29 – 94 mg kg^-1).

Compost application significantly improved soil acid-base conditions and fertility. One year after its incorporation (PC t₁, Table 1), compost greatly increased soil pH (pH_H2O 5.3 ± 0.2), C_T and N_T content (7.5 ± 1.0% and 0.5 ± 0.1%, respectively), as well as P_Olsen, whose content increased nearly 100-fold on average (285 ± 44 mg kg^-1). The application of compost also notably increased CEC_e up to 40.7 ± 2.7 cmol_c kg^-1 and completely altered the composition of the exchange complex. Exchangeable acidity (H⁺ and Al³⁺), which was dominant in the unamended soil, was replaced by Ca²⁺ as the dominant cation (37.7 ± 2.3 cmol_c kg^-1), accompanied by smaller amounts of Mg²⁺ and K⁺ (2.9 ± 0.6 and 1.0 ± 0.2 cmol_c kg^-1, respectively). As a result, base saturation exceeded 90% (on average). The application of compost did not alter the pseudo-total concentration of Cu (Cu_T) but drastically reduced its availability (Cu_RA = 3.8 ± 1.6 mg kg^-1). In addition, an important increase was observed in pseudo-total Mn, Zn, and Pb (Mn_T, Pb_T, and Zn_T), although their availability was very low (Mn_RA = 7.6 ± 1.7, Zn_RA = 4.6 ± 1.1 mg kg^-1) or undetectable (Pb_RA).

At the same time, the application of technosols also significantly improved soil fertility parameters, although with some differences compared to compost. Thus, both mixtures showed a more pronounced increase in pH (pH_H2O 7.0 ± 0.2 and 6.8 ± 0.2, in PT1 t₁ and PT2 t₁, respectively), in CEC_e (41.4 ± 2.2 and 58.5 ± 2.1 cmol_c kg^-1, PT1 t₁ and PT2 t₁) and exchangeable bases, particularly Ca²⁺ (36.6 ± 1.7 and 53.9 ± 1.9 cmol_c kg^-1, PT1 t₁ and PT2 t₁) and K⁺ (1.7 ± 0.2 and 2.0 ± 0.2 cmol_c kg^-1, PT1 t₁ and PT2 t₁). In contrast, the increase in C_T (3.7 ± 0.4 and 4.4 ± 0.4 %, PT1 t₁ and PT2 t₁), N_T (0.25 ± 0.04 and 0.32 ± 0.03 %, PT1 t₁ and PT2 t₁), and P_Olsen (187 ± 45 and 132 ± 16 mg kg^-1, PT1 t₁ and PT2 t₁) was lower than in compost treated soil. Similarly to compost, technosols increased Mn_T, Zn_T (PT2 only), and Pb_T content, although this increase was more pronounced for Mn_T (1663 ± 82 and 1926 ± 93 mg kg^-1, PT1 t₁ and PT2 t₁) and less so for Pb_T and Zn_T. A relative decrease in the concentration of Cu_T was also observed (267 ± 32 and 421 ± 83 mg kg⁻¹, PT1 t₁ and PT2 t₁). Similar to the results observed with compost, the availability of PTEs was reduced to very low levels with the application of technosols.

Table 1. Main soil physicochemical properties, metal contents (pseudo-total) and PTE availability (NaNO₃-extractable) at the trial plots before (t₀) and after one year (t₁) of being amended with compost (PC) and technosols (PT1, PT2).

	PC			PT1			PT2
	t0	t1		t0	t1		t0	t1
pH H2O	3.13 ± 0.03	5.29 ± 0.24		2.81 ± 0.03	6.98 ± 0.20		2.96 ± 0.04	6.77 ± 0.19
pH KCl	3.02 ± 0.03	5.11 ± 0.23		2.62 ± 0.03	6.72 ± 0.16		2.69 ± 0.04	6.73 ± 0.17
CT (%)	0.39 ± 0.09	7.50 ± 1.02		0.69 ± 0.10	3.68 ± 0.43		0.50 ± 0.05	4.45 ± 0.37
NT (%)	0.1 ± 0.01	0.52 ± 0.10		0.10 ± 0.01	0.25 ± 0.04		0.10 ± 0.00	0.32 ± 0.03
POlsen (mg kg-1)	2.0 ± 0.0	285.1 ± 44.1		< loq	186.6 ± 45.2		< loq	132.0 ± 16.2
CECe (cmolc kg-1)	23.7 ± 5.6	40.7 ± 2.7		29.3 ± 9.7	41.4 ± 2.2		19.5 ± 3.8	58.5 ± 2.1
H+	9.1 ± 1.8	-		15.3 ± 4.2	-		9.9 ± 1.7	-
Al3+	9.6 ± 2.0	0.4 ± 0.4		8.4 ± 2.2	1.0 ± 0.9		7.0 ± 1.1	< loq
Ca2+	3.0 ± 1.0	37.7 ± 2.3		2.8 ± 1.0	36.6 ± 1.7		3.0 ± 0.9	53.9 ± 1.9
K+	< loq	1.0 ± 0.2		< loq	1.7 ± 0.2		< loq	2.0 ± 0.2
Mg2+	2.0 ± 0.8	2.9 ± 0.6		1.2 ± 0.5	2.6 ± 0.2		1.2 ± 0.4	2.5 ± 0.1
Pseudo-total concentrations
(g kg-1)
AlT	45.7 ± 0.8	38.4 ± 1.2		40.0 ± 1.0	34.8 ± 0.6		36.9 ± 0.6	40.9 ± 1.2
FeT	108.1 ± 2.7	67.0 ± 4.2		108.2 ± 2.6	58.1 ± 6.0		103.7 ± 1.6	56.6 ± 7.5
(mg kg-1)
MnT	848.4 ± 15.0	1127.4 ± 39.7		787 ± 15.5	1662.6 ± 81.7		828.2 ± 19.1	1925.7 ± 93.5
CuT	567.7 ± 34.0	499.8 ± 33.6		622.7 ± 47.7	267.3 ± 32.3		629.8 ± 42.8	420.6 ± 83.4
ZnT	131.6 ± 4.6	339.6 ± 38.9		131.2 ± 11.6	140.6 ± 16.3		117.3 ± 4.1	191.6 ± 8.5
NiT	51.3 ± 6.4	65.5 ± 17.0		47.5 ± 8.6	31.8 ± 2.1		40.2 ± 7.1	33.0 ± 1.9
CrT	102.9 ± 5.9	92.2 ± 4.8		94.3 ± 3.1	79.2 ± 18.1		89 ± 2.8	74.4 ± 6.5
PbT	16.2 ± 0.4	53.5 ± 7.2		17.7 ± 0.8	33.2 ± 2.7		17.6 ± 0.3	43.7 ± 3.4
CdT	1.5 ± 0.1	1.0 ± 0.2		1.7 ± 0.1	1.5 ± 0.1		1.6 ± 0.1	0.8 ± 0.2
NaNO3-extractable (mg kg-1)
MnRA	93.9 ± 8.0	7.6 ± 1.7		52.1 ± 7.5	3.2 ± 1.5		28.8 ± 2.1	2.5 ± 1.2
CuRA	51.5 ± 9.5	3.8 ± 1.6		107.8 ± 13.8	1.4 ± 0.5		67.3 ± 5.4	1.9 ± 0.8
ZnRA	19.1 ± 2.7	4.6 ± 1.1		36.6 ± 8.9	0.8 ± 0.0		20.2 ± 2.7	0.9 ± 0.6
NiRA	8.4 ± 1.1	0.9 ± 0.3		13.1 ± 2.1	0.4 ± 0.2		7.7 ± 0.9	0.3 ± 0.1
CrRA	0.1 ± 0.0	< loq		0.5 ± 0.1	< loq		0.3 ± 0.0	< loq
CdRA	0.1 ± 0.0	< loq		< loq	< loq		0.1 ± 0.0	< loq
loq, limit of quantification: 0.01 cmolc kg-1 Al3+, K+; 0.04 cmolc kg-1 Na+; 0.01 mg kg-1 Cr, Pb, Cd; 0.2 mg kg-1 POlsen;
excahgeable Na+ and PbRA always < loq

Table 1. Main soil physicochemical properties, metal contents (pseudo-total) and PTE availability (NaNO₃-extractable) at the trial plots before (t₀) and after one year (t₁) of being amended with compost (PC) and technosols (PT1, PT2).

	PC			PT1			PT2
	t0	t1		t0	t1		t0	t1
pH H2O	3.13 ± 0.03	5.29 ± 0.24		2.81 ± 0.03	6.98 ± 0.20		2.96 ± 0.04	6.77 ± 0.19
pH KCl	3.02 ± 0.03	5.11 ± 0.23		2.62 ± 0.03	6.72 ± 0.16		2.69 ± 0.04	6.73 ± 0.17
CT (%)	0.39 ± 0.09	7.50 ± 1.02		0.69 ± 0.10	3.68 ± 0.43		0.50 ± 0.05	4.45 ± 0.37
NT (%)	0.1 ± 0.01	0.52 ± 0.10		0.10 ± 0.01	0.25 ± 0.04		0.10 ± 0.00	0.32 ± 0.03
POlsen (mg kg-1)	2.0 ± 0.0	285.1 ± 44.1		< loq	186.6 ± 45.2		< loq	132.0 ± 16.2
CECe (cmolc kg-1)	23.7 ± 5.6	40.7 ± 2.7		29.3 ± 9.7	41.4 ± 2.2		19.5 ± 3.8	58.5 ± 2.1
H+	9.1 ± 1.8	-		15.3 ± 4.2	-		9.9 ± 1.7	-
Al3+	9.6 ± 2.0	0.4 ± 0.4		8.4 ± 2.2	1.0 ± 0.9		7.0 ± 1.1	< loq
Ca2+	3.0 ± 1.0	37.7 ± 2.3		2.8 ± 1.0	36.6 ± 1.7		3.0 ± 0.9	53.9 ± 1.9
K+	< loq	1.0 ± 0.2		< loq	1.7 ± 0.2		< loq	2.0 ± 0.2
Mg2+	2.0 ± 0.8	2.9 ± 0.6		1.2 ± 0.5	2.6 ± 0.2		1.2 ± 0.4	2.5 ± 0.1
Pseudo-total concentrations
(g kg-1)
AlT	45.7 ± 0.8	38.4 ± 1.2		40.0 ± 1.0	34.8 ± 0.6		36.9 ± 0.6	40.9 ± 1.2
FeT	108.1 ± 2.7	67.0 ± 4.2		108.2 ± 2.6	58.1 ± 6.0		103.7 ± 1.6	56.6 ± 7.5
(mg kg-1)
MnT	848.4 ± 15.0	1127.4 ± 39.7		787 ± 15.5	1662.6 ± 81.7		828.2 ± 19.1	1925.7 ± 93.5
CuT	567.7 ± 34.0	499.8 ± 33.6		622.7 ± 47.7	267.3 ± 32.3		629.8 ± 42.8	420.6 ± 83.4
ZnT	131.6 ± 4.6	339.6 ± 38.9		131.2 ± 11.6	140.6 ± 16.3		117.3 ± 4.1	191.6 ± 8.5
NiT	51.3 ± 6.4	65.5 ± 17.0		47.5 ± 8.6	31.8 ± 2.1		40.2 ± 7.1	33.0 ± 1.9
CrT	102.9 ± 5.9	92.2 ± 4.8		94.3 ± 3.1	79.2 ± 18.1		89 ± 2.8	74.4 ± 6.5
PbT	16.2 ± 0.4	53.5 ± 7.2		17.7 ± 0.8	33.2 ± 2.7		17.6 ± 0.3	43.7 ± 3.4
CdT	1.5 ± 0.1	1.0 ± 0.2		1.7 ± 0.1	1.5 ± 0.1		1.6 ± 0.1	0.8 ± 0.2
NaNO3-extractable (mg kg-1)
MnRA	93.9 ± 8.0	7.6 ± 1.7		52.1 ± 7.5	3.2 ± 1.5		28.8 ± 2.1	2.5 ± 1.2
CuRA	51.5 ± 9.5	3.8 ± 1.6		107.8 ± 13.8	1.4 ± 0.5		67.3 ± 5.4	1.9 ± 0.8
ZnRA	19.1 ± 2.7	4.6 ± 1.1		36.6 ± 8.9	0.8 ± 0.0		20.2 ± 2.7	0.9 ± 0.6
NiRA	8.4 ± 1.1	0.9 ± 0.3		13.1 ± 2.1	0.4 ± 0.2		7.7 ± 0.9	0.3 ± 0.1
CrRA	0.1 ± 0.0	< loq		0.5 ± 0.1	< loq		0.3 ± 0.0	< loq
CdRA	0.1 ± 0.0	< loq		< loq	< loq		0.1 ± 0.0	< loq
loq, limit of quantification: 0.01 cmolc kg-1 Al3+, K+; 0.04 cmolc kg-1 Na+; 0.01 mg kg-1 Cr, Pb, Cd; 0.2 mg kg-1 POlsen;
excahgeable Na+ and PbRA always < loq

significance levels:
nse p > 0.05
* p < 0.05
** p < 0.01
*** p < 0.001

3.2. Medium-to-Long-Term Evolution of Soil Physicochemical Properties

The temporal dynamics of soil parameters were studied using GLS models. Table 2 shows the main statistical descriptors, and Figure 1 and Figure 2 illustrate the dynamics of most characteristic properties over a ten-year period. Pb (Pb_T, Pb_RA) and Cd (Cd_T, Cd_RA) and pH_KCl did not show any meaningful change over time, thus they were excluded from the evolutionary analysis.

In the untreated controls (UNT), there was virtually no temporal variation in pH_H2O or in the (pseudo-)total content of most of the elements analyzed (C_T, N_T, Al_T, Cu_T, Pb_T, Cd_T). There were significant changes (p < 0.05) for P_Olsen, CEC_e, and the exchangeable cations (K+, Mg²⁺, Ca²⁺, and Al³⁺), as well as for Fe_T, Mn (Mn_T, Mn_RA), Zn (Zn_T, Zn_RA), and Cu_RA (Table 2). P_Olsen (Figure 1D), exchangeable K⁺, and Fe_T increased slightly over time, while the rest of the parameters tended to decrease, sometimes abruptly (Figure 2). Overall, the UNT soils acquired certain nutrients and became enriched in Fe_T (p < 0.001), while indicators related to particle retention and solubility – such as CEC_e (Figure 2A,B) and the labile fractions (Figure 2C,D), decreased progressively (p < 0.001) along with some total PTEs, such as Zn_T (p < 0.001) and Mn_T (p = 0.020).

The evolution of the treated soils was more complex, with significant effects of the various factors (Amendment, Time, and Plantation) and their interactions. Most of the variation in soil parameters was explained independently by Amendment and Time factors (Table 2). The Amendment factor explained significant differences (p < 0.01) in all parameters analyzed, except for exchangeable Mg²⁺, while Time was particularly relevant (p < 0.001) for C_T, N_T, P_Olsen, CEC_e, exchangeable Mg²⁺ and Ca²⁺, and Mn_T. C_T and N_T increased overtime in all treatments (Figure 1B,C), while the remaining parameters showed a decrease (Figure 1D and Figure 2). Other parameters showed either insignificant/ negligible temporal change (pH_H2O, exchangeable K⁺ and Al³⁺, Al_T, Fe_T, Cu_T, Mn_RA), or poor correspondence between model predictions and observed variability (Zn_T, Cu_RA, Zn_RA).

The reciprocal modulation between the Amendment and Time factors – A*T interaction – was particularly significant (p < 0.001) for N_T, CEC_e, exchangeable cations, and Al_T (Table 2). Statistical scrutiny of this interaction reveals differences in the temporal dynamics among amendments. N_T increased more rapidly in PC than in the technosols-amended plots (Figure 1C), remaining around 0.50% during the first three years and doubling over the following seven years. In contrast, in both technosols, N_T remained around 0.35% during the first six years and increased by only 0.10% during the t₆ – t₁₀ period. CEC_e showed a relatively faster decline in PT2 – which started from higher initial values due to the Amendment effect – than in PC and PT1 (Figure 2A). The overall downward trend, associated with Time, appears to converge toward values around 20 cmolc kg^-1 after 10 years (t₁₀). Exchangeable cations followed approximately the same trends as CEC_e; however, Mg²⁺ remained relatively stable throughout the series (Mg²⁺ ≈ 2.00 cmolc kg^-1; Figure 2B), as well as K⁺ in PC (K⁺ ≈ 1.00 cmolc kg^-1; data not shown). The statistically significant decrease in Al_T was only observable in PT2 between t₃ and t₁₀ (data not shown). For the remaining parameters, the A*T interactive effect had little or no influence. C_T and Cu_RA, key parameters in this study, were A*T influenced to some extent (p < 0.01) resulting in a faster increase in C_T in PC (Figure 1A) and minimal fluctuations in Cu_RA (Figure 2C). C_T increased, on average, by 4.4% in PC over the whole period, and by 1.0 – 2.0% in technosols-amended soils starting at t₆. Considering the N_T trends described above, the C:N ratio decreased by 4.7 points in PC, by 2.3 in PT1, and by 1.8 in PT2, eventually converging to values close to 12. Cu_RA decreased slightly in both mean concentration and dispersion in the technosols-amended plots from t₁ to t₃. It subsequently increased from t₃ onward, reaching lower levels in PC and higher levels especially in PT2. Nevertheless, Cu_RA remained close to 1.0 mg kg^-1 throughout the trial.

The plants (Plantation factor) had statistically significant effects on various parameters (Table 2). Although, in general, these effects were of lesser magnitude or were closely associated with other factors, particularly Time. Thus, P*T interaction was statistically relevant (p < 0.001) for C_T and N_T, CEC_e, exchangeable Mg²⁺ and Ca²⁺, Cu_T, Cu_RA and Zn_T. Figure 1B shows these distinct dynamics on C_T. The A. capillaris (AGR) planted plots tended to show early (t₁ – t₃) increases in C_T, whereas in S. viminalis (SAL the increases in C_T started later but showed a sharper rise. Paradoxically, the control without planting (UNP) also experienced a general increase in C_T; however, this increase was more delayed and less consistent than in AGR or SAL. By the end (t₁₀), C_T tended to converge in PC plots toward values around 12% regardless of the Plantation treatment. A similar trend was observed in PT1, where UNP and AGR converged toward the same C_T value (5%) at t₁₀. Nevertheless, in SAL (PT1 and PT2), as well as in AGR (PT2), C_T eventually diverged from UNP toward 6%. N_T evolved similarly to C_T (Figure 1C), although it increased more rapidly, resulting in a decrease in the C:N ratio parallel to the increase in C_T starting approximately at t₁ in AGR, at t₃ in SAL, and at t₆ in UNP, reaching values close to 12. On the other hand, CEC_e decreased (Figure 2A), more rapidly and from higher initial values in AGR, especially in PC and PT1, finally converging toward UNP levels around t₁₀. The trend was practically identical for exchangeable Ca²⁺, since it accounted for approximately 80% of the CEC_e, and it was similar for exchangeable Mg²⁺ (Figure 2B). Introduced vegetation was associated with only marginally higher Cu_RA and lower Zn_T values than in UNP plots. Complex effects detected for these parameters should be interpreted cautiously because their models had relatively low coefficients of determination (R² < 0.4).

Finally, the main statistical differences associated with the three-way interaction (A*P*T, Table 2) were the variations in C_T and N_T. The model indicates that Time was the most important factor for C_T, followed by Plantation, while Amendment acted primarily as a modulator. Figure 1B summarizes this pattern, showing that the predominant trend was the progressive accumulation of C_T in all soils – more intense or earlier in those with introduced vegetation, greater in PC than in technosols-amended plots, and at different rates depending on the amendment and planting regime. N_T shared with C_T the predominance of the Time and Plantation effects, although it differed in the A*P interaction which was equal in magnitude to the isolated Plantation effect. Other parameters also showed three-way interaction effects, but the separate single factors or two-way interactions more adequately explained their variation.

3.3. Multivariate Analysis

Principal Component Analysis (PCA) accounted for 73,7% of the total variance with the first three principal components (CP), clearly identifying the Time and Amendment factors on soil response and evolution; from this perspective Plantation was less relevant. The first component (CP1, 42,3% of the total variance) was positively associated with variables indicative of fertility (pH, CEC_e, exchangeable Ca²⁺ and K⁺, C_T and N_T, and P_Olsen), as well as Mn_T and Zn_T (Figure 3A,C). On the other hand, the negative loading of CP1 was associated with total metals (Fe_T, Cu_T) and all readily available PTEs, which were characteristic of the untreated mine soil (Table 1). The second component (CP2, 17,3% of the total variance) was positively associated with cation exchange capacity (CEC_e) and exchangeable cations, especially with exchangeable Mg²⁺, as well as P_Olsen, Mn_T, and all readily available PTEs. CP2 was negatively associated with Time included as a supplementary variable (Figure 3A). CP2 arranges the samples chronologically and reflects soil evolution, with a progressive loss of the most mobile elements and a clear distinction between the early years (t₀ – t₃), located in the positive range of the axis, and the later years (located in the negative range). Plotting the samples on the CP1 – CP2 plane (Figure 3B) places 2011 soils (prior to treatment) at the most negative end of CP1 and those from 2012 (PC, PT1, and PT2) at the most positive end, reflecting how the amendments rapidly changed the initial system. The third component (CP3, 14,1% of the total variance) was positively associated with C_T, N_T, P_Olsen, and Zn_T, and negatively with pH, CEC_e, exchangeable Ca²⁺ and K⁺, Al_T, and Mn_T (Figure 3C). Sample projection onto the CP1 – CP3 plane (Figure 3D) shows a clear distinction between soil amendments, and it is interpreted as a gradient of organic matter, fertility, and introduced PTEs.

4. Discussion

4.1. Amendments Rapidly Transform Heavily Degraded Soil Into a Functional System

The application of soil amendments to soils affected by mining is a widely used strategy to reduce toxicity from PTE, improve fertility, and facilitate plant establishment in aided phytostabilization systems. In acidic soils, such as those studied in this work, the use of alkalizing materials is particularly important, since neutralizing extreme acidity reduces metal solubility, enhances nutrient availability, and improves soil functionality [10,25]. This initial correction of soil limitations is, in many cases, a prerequisite for the establishment of a functional phytomanagement system [15,18,48,61,62].

The results of this study confirm this function of soil amendments. The mine soil (UNT) exhibited extreme acidity, low levels of C_T, N_T and P_Olsen, low CEC_e, and high concentrations of total and readily available metals and PTE, especially Cu (Cu_T, Cu_RA). Following the application of compost or technosols, the soil system shifted toward neutrality, with higher carbon and nutrient (N, P, Ca, Mg, K) contents, an increase in CEC_e, and a drastic reduction in the availability of Cu and other PTEs. Multivariate analysis summarizes this transformation in a gradient associated with CP1, which clearly separated the acidic, metal-rich mine soil from the amended soils, which were more fertile and less phytotoxic. This finding is consistent with other studies on soils and waters affected by acid mine drainage, in which pH neutralization was associated with a decrease in the mobility of PTEs and a general improvement in soil properties [5,7,63]. Acidity adjustment directly affects the speciation and mobility of metals, as well as nutrient availability, soil physical structure, and microbiological activity [25,26,61,64]. In the present study, the decrease in Cu_RA was particularly marked, with levels in the amended soils two or three orders of magnitude lower than those observed in UNT. Conversely, Cu_T decreased by only half at most. This indicates that the amendments not only led to a relative dilution of the total content of Cu, owing to their low content of this element, but also substantially altered its reactivity and environmental risk.

4.2. The different Amendments Lead to Distinct Initial Trajectories Depending on Their Composition

Although all amendments clearly improved soil conditions, the results show that their functional improvement was unique. In fact, the multivariate analysis identified a gradient associated with the type of amendment, determined by differences in specific elements, pH and CEC_e. This pattern indicates that the composition of the applied materials exerted a strong and lasting influence on the system’s properties, even when the different amendments partially shared the use of certain components, such as sewage sludge.

The compost treatment (PC) was characterized by higher levels of C_T, N_T, P_Olsen, and, to a lesser extent, exchangeable Mg²⁺. In this regard, its soil-amending effect was primarily associated with a higher input of organic matter and nutrients, which is consistent with the composition of composts produced from similar organic waste, which are generally rich in C, N, P, and other macronutrients [23,26]. The 7 – 8% C_T level initially achieved in the PC plots (t₁) falls within the bioclimatic equilibrium range of natural soils in the area [65]. This content in organic matter is uncommon in the untreated mine soils of the area, but it could be maintained due to the enhanced nutrient status via vegetation inputs in the long-term.

In addition, in this study an increase in Zn_T (and, to a lesser extent, Mn_T and Pb_T) was observed following compost application. Fagnano et al. [18] noted that in agricultural soils sewage sludge addition can contribute secondary PTEs, particularly Zn, Cd, or Pb, but the associated risk depends on specific agricultural practices, such as application rate, frequency of applications, pH control, and land use. Furthermore, composting promotes the stabilization of organic matter, reduces the proportion of dissolved organic matter, and can decrease the risk of PTE mobilization through soluble complexation. This stabilization is particularly relevant in soils contaminated with Cu, Pb, or As, where immature organic materials can temporarily increase PTE mobility [66,67]. Here, PTE availability remained at low levels suggesting that the elements introduced by the soil amendments were stabilized, likely associated with organic matter, colloidal surfaces, or low-solubility phases. Therefore, in this context – nonagricultural soils with controlled pH and very low PTE availability – the immediate risk appears limited, but should continue to be monitored [68].

The technosols applied in PT1 and PT2, meanwhile, exhibited a stronger alkalizing effect than compost, reaching significantly higher levels of pH, CEC_e, as well as exchangeable Ca²⁺ and K⁺. These effects are related to the use of basic/alkaline residues and reactive mineral materials, including biomass ash, carbonates bound to paper mill sludge, and Fe or Al (hydr)oxides, capable of neutralizing acidity. The dissolution of basic salts (such as CaCO₃) or alkaline hydroxides (such as KOH) provides nutrients, raises the pH, and increases the negative charge of colloidal amphoteric surfaces, which increases cation retention and contributes to the immobilization of PTEs [17,21,69]. Other authors reported similar pH increases following the application of biomass ash (or limestone amendments) to PTE-contaminated soils, correlating this with a decrease in the bioavailable fraction due to precipitation, adsorption, and complexation processes [22,24]. Furthermore, Al and Fe (hydr)oxides provide highly reactive mineral surfaces that can promote the sorption and retention of PTEs and macronutrients [7,21,70].

However, the increased pH of the technosols was accompanied by an increase in total concentrations of some elements, particularly Mn and Al (PT2). This result suggests a contribution from the starting materials, likely linked to ash and other components rich in these elements (e.g., aluminum industry waste in PT2). Even though their readily available fractions remained very low, the overall increase indicates that the recovery of industrial waste in technosols must be carefully evaluated. Complex amendments can very effectively improve fertility and reduce the mobility of PTEs, but they can also introduce secondary elements whose stability will depend on changes in pH, organic matter, redox conditions, and the biological functioning of the system [10,15,26].

The combination of organic and inorganic materials in technosols offers potential advantages, nevertheless. Organic materials provide binding sites for PTEs and sources of carbon and nutrients, while mineral components can promote flocculation, the stabilization of organic matter, and the immobilization of PTEs through oxides, hydroxides, or organomineral complexes [10,15,26]. Some authors [27,67] suggested that best organic amendments for PTE-contaminated soils are those that mature in situ, allowing PTE associated with low-molecular-weight molecules (especially Cu) to be incorporated into increasingly larger structures. However, there is a risk of aqueous-phase mobility of these low-molecular-weight molecules, especially when they are rich in carboxyl groups [69], or transfer to the food chain through bioaccumulation in plant biomass [18,68]. Technosols, in contrast to compost, were designed so that the inorganic components would modulate the stabilization of the organic components in situ. Cations with high charge-to-radius ratios, such as Ca²⁺, Al³⁺, and Fe³⁺, promote the flocculation of organic matter [69,71]. Al and Fe cations and (hydr)oxides can participate in the formation of organometallic complexes that stabilize organic matter, making it more recalcitrant, while also immobilizing PTEs and increasing the stability of aggregates [65,72,73,74]. These synergistic effects were not obvious in the short term, but they were expected in the long term.

4.3. Remediation Processes Continue up to Ten Years

One of the most significant findings of this trial is the confirmation that the amended soils did not reach a steady state after the initial intervention but continued to evolve throughout the entire follow-up period. This observation is particularly important because, as noted in literature, medium-to-long-term field studies remain very scarce compared to laboratory, greenhouse, or short-term experiments [10,15,24,37,38]. The trends observed in this trial confirm that the effectiveness of aided phytoremediation cannot be assessed solely based on the early effects of the amendments.

The multivariate analysis revealed a temporal dimension (CP2) characterized by the progressive decline of certain nutrients, exchangeable cations, and bioavailable PTEs (which remained low throughout the experiment). In UNT, the decrease in Cu_RA and other PTEs labile pools can be interpreted as natural local attenuation, likely associated with leaching processes favored by the high soil acidity and rainfall in the study area. However, this apparent local decrease does not imply effective environmental recovery, as the mobilization of PTEs can lead to the transfer of contamination to surface or groundwater. This limitation is one of the main weaknesses of natural attenuation processes in mining systems: they can partially reduce the local concentration of contaminants, but over long-time scales and with a persistent risk of dispersion [7].

In the treated soils, the temporal evolution was marked by a decline in CEC_e, exchangeable bases, P_Olsen, and, in some cases, Mn_T, which converged toward the same point in all treatments. This trend can be explained by a combination of leaching processes, plant uptake, transformation of the reactive phases of the amendments, and a gradual adjustment toward a new soil equilibrium. The PT2 soil showed a more rapid decline in CEC_e and bases, likely because it started from higher initial values. Similarly, P_Olsen tended to converge from initially very high values, especially in PC. This proportionality between initial value and rate of change suggests that the long-term evolution trajectories across the different treatments are defined by different rates within similar mechanisms, rather than by clearly distinct mechanisms. Some of the properties induced by the amendments would undergo gradual attenuation toward a common equilibrium, consistent with the edaphoclimatic conditions of the area.

The most representative natural soils in the area (Haplic Umbrisol, WRB, 2015) are generally acidic (pH_H2O 4.5 – 5.5), with low CEC_e (bellow 10 cmol_c kg^-1), a desaturated exchange complex dominated by Al³⁺, and very low chemical fertility (especially P and K) [75]. After ten years, the pH of the treated soils remained around neutral, the CEC_e tended toward values bellow 20 cmol_c kg^-1, the exchangeable complex was still dominated by bases, and the P_Olsen content remained high (especially in PC, 133 ± 6 mg P_Olsen kg^-1). This indicates that, although there was a gradual loss of alkalinity, basic cations, and nutrients, the soil-improving effect of the amendments persists in the long term. The system may continue to evolve toward a gradual loss of bases, but as long as the pH remains buffered, the equilibrium value of CEC_e should remain reliably high.

4.4. Introduced Vegetation Influences Recovery, Although Its Effect Is Partially Masked

The Plantation effect was statistically significant for various edaphic variables, although, in general, it was of lesser magnitude than the effect Amendment or Time. This hierarchy is consistent with an initially hostile system, in which the chemical correction imposed by the amendments is the main driver of restoration. However, plants clearly modulated the evolution of C_T, N_T, P_Olsen, CEC_e, indicating the significant role of vegetation dynamics in post-soil-amendment biogeochemical evolution.

When interpreting the Plantation effect, it should be noted that the unplanted plots (UNP) were colonized early on by adventitious herbaceous plants. This colonization was favored by the initial fertility surge generated by the amendments. Furthermore, after the initial period (t₃), the agronomic practice of weed control recommended in phytomanagement systems [14] ceased. Consequently, the UNP treatment does not represent soil without vegetation, but rather a scenario of spontaneous colonization, which reduces the statistical contrast between planting treatments.

One of the most sensitive variables to the interaction among amendment, vegetation, and time was C_T. In PC, a sustained accumulation of C_T was observed, with an apparent internal convergence toward values close to 12%. This dynamic might be explained by the greater initial contribution of nutrient rich organic matter from the compost, with the subsequent recycling and incorporation of carbon via necromass (plant and microbial). In the early years, differences in C_T in the PC plots were associated with the driving effect of A. capillaris in AGR plots, whereas the gradual introduction of adventitious herbaceous plants in the UNP and SAL plots, together with the development of S. viminalis, favored convergence among the different vegetation covers toward the end of the experiment. In PT1 and PT2, net increases in C_T occurred later and were more moderate, with values close to 5% in plots dominated by herbaceous plants (UNP and AGR) and slightly higher in those planted with S. viminalis (SAL). This suggests that the combined presence of woody and herbaceous vegetation may promote greater carbon accumulation in the medium term. The parallel evolution of C_T and N_T, and the convergence of the C:N ratio toward values close to 12 suggest a progressive biogeochemical stabilization of organic matter consistent with the dynamics of organic matter in natural soils of the region [65].

The faster accumulation of organic matter in PC plots could indicate greater early biological activity and a greater contribution of microbial necromass to stabilized carbon. This interpretation is consistent with the results of Touceda-González et al. [48], who observed (in the same PC plots used in this trial) a marked increase in enzymatic activity at the start of the experiment, including high urease activity, indicative of an intensification of the N cycle. Although they assumed that compost introduced microorganisms and enzymes not previously present in the mining waste, enzymatic activity remained stable or even increased over time (3 years), raising the hypothesis that the addition of compost facilitated the establishment of stable and active microbial populations in the mine soils. Tica et al. [25] also noted the sensitivity of microbiological activity to the availability of PTEs and nutrients, which reinforces the idea that the initial soil improvement may have rapidly activated key microbial processes. The relative immaturity of the sludge incorporated into the technosols may have resulted in a less stable microbial community, with poorer biogeochemical recycling capacity, reduced potential for symbiosis, or a higher presence of phytopathogenic microorganisms in PT1 and PT2. Further specific analysis would clarify whether this apparent buildup of organic carbon is solely a consequence of greater influx of fresh (and labile) C or reflects effective stabilization in recalcitrant forms [76,77].

The dynamics of CEC_e and P_Olsen also suggest an interaction between plant activity, nutrient recycling, and system losses. The Plantation factor (and its interactions) was more significant than the A*T interaction in P_Olsen, suggesting a significant modulation of phosphorus dynamics by vegetation. In contrast, Plantation is a secondary factor in CEC_e. However, in AGR plots an early response was observed, consistent with rapid root colonization of introduced A. capillaris, initial nutrient mobilization, and subsequent uptake during growth. During the first few years, part of the biomass was removed for a study of plant productivity [48], which may have limited the return of nutrients to the soil and contributed to the subsequent decline in P_Olsen and exchangeable bases. Starting in the early years, the incorporation of spontaneous biomass may have promoted some internal recycling, although this was not necessarily sufficient to compensate for losses due to leaching, occlusion, or export via plant tissues. Furthermore, Touceda-González et al. [48] observed that nutrient concentrations in A. capillaris and in the leaves of S. viminalis tended to decrease over time, likely reflecting an initial surge in nutrient uptake followed by a subsequent phase of growth and dilution due to an increase in structural elements.

Differences between compost and technosols may also have directly affected plant development. Although the technosols provided greater acid neutralization and a higher initial cation exchange capacity (CEC_e), their pH – which was slightly alkaline – may have been suboptimal for the acid-adapted species used in the experiment. Likewise, a nutritional imbalance caused by excess Ca²⁺ or potentially phytotoxic levels of Mn may have limited plant establishment in PT1 and PT2. Nutrient imbalance may reduce water-use efficiency in some plants [61]. Although not a limiting factor, P_Olsen in technosols-amended plots was lower than in PC, perhaps reflecting microsite-level retention of available P on highly reactive surfaces of ash, Fe or Al (hydr)oxides [70]. PC also exhibited more favorable physical properties, including lower bulk density and higher water-holding capacity [78], which may have enhanced plant resilience to water stress. This aspect is relevant, as water efficiency is considered critical for the sustainability of phytomanagement systems [14,15,61].

5. Conclusions

This study confirms that aided phytostabilization can transform a hyperacidic, infertile, and phytotoxic mine soil into a functional soil system for at least a decade. The soil amendments rapidly corrected the acidity, increased fertility, and immobilized potentially toxic PTEs in situ, thereby enabling vegetation establishment and the subsequent evolution of the soil-plant system. These changes are consistent with the objectives of phytomanagement to reduce environmental risk, promote ecological recovery, and create conditions for the provision of ecosystem services, such as soil stabilization, carbon sequestration, and potential biomass production.

The results also suggest that under the same climatic conditions the dynamics of different amendments tend to converge toward a common long-term equilibrium. Their short-to-medium-term evolution differs, nevertheless, and distinct strategies led to distinct soil trajectories. From an applied perspective, therefore, the choice of amendment should align with the priority short- and long-term objectives. Compost is suitable when the goal is to increase nutrient fertility, soil organic carbon, and biological activity early on. Designed technosols offer greater pH control, higher cation exchange capacity, and more diverse mechanisms for PTE fixation; therefore, they can be useful in scenarios where acid neutralization and short-term PTE stabilization are priorities. However, the use of waste materials – both in compost production and in technosols formulation – requires careful monitoring regarding secondary PTEs (Al, Mn, Pb, Zn) and their long-term behavior.

The study also emphasizes that the effectiveness of an intervention should not be evaluated solely in the short term. The amendments produced rapid improvements, but the system continued to evolve over a decade, with progressive losses of certain cations and nutrients, differential carbon accumulation, and changes in Cu availability. Therefore, phytomanagement programs should incorporate long-term monitoring plans, focusing not only on plant survival/growth or total PTE concentrations, but also on functional soil indicators such as pH, CEC_e, exchangeable bases, C_T, N_T, P_Olsen, and readily available fractions of PTEs.

Finally, the differences observed between soil treatments and plantation types indicate that the mechanisms controlling the system’s evolution are complex and partially dependent on the interaction among materials, vegetation, and local conditions. Future work should delve deeper into the dynamics PTE, C, N, and P fractions, the mineralogy of reactive phases, the formation of organomineral complexes, and microbial activity. To comprehensively evaluate the phytomanagement system over the long term, it would also be necessary to expand monitoring to other key compartments – particularly runoff and leachate, as well as the plant ionome – in order to assess the system’s environmental stability and the actual potential for biomass recovery.