Next Article in Journal
Effects of Conservation Tillage and Nitrogen Management on Yield, Grain Quality, and Weed Infestation in Winter Wheat
Previous Article in Journal
Method and Optimization of Key Parameters of Soil Organic Matter Detection Based on Pyrolysis Coupled with Artificial Olfaction
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Agronomy
Volume 15
Issue 7
10.3390/agronomy15071741
agronomy-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Rare Earth Elements in Tropical Agricultural Soils: Assessing the Influence of Land Use, Parent Material, and Soil Properties

by
Gabriel Ribeiro Castellano
1,*,
Juliana Silveira dos Santos
2,
Melina Borges Teixeira Zanatta
1,
Rafael Souza Cruz Alves
3,
Zigomar Menezes de Souza
4,
Milton Cesar Ribeiro
3 and
Amauri Antonio Menegário
1,*
1
Environmental Studies Center, São Paulo State University (CEA/UNESP), Rio Claro 13506-900, SP, Brazil
2
Instituto Nacional da Mata Atlântica (INMA), Museu de Biologia Professor Mello Leitão, Av. José Ruschi, 4-Centro, Santa Teresa 29650-000, ES, Brazil
3
Laboratório de Ecologia Espacial e Conservação, São Paulo State University (LEEC/UNESP), Rio Claro 13506-900, SP, Brazil
4
School of Agricultural Engineering, Universidade Estadual de Campinas (FEAGRI/UNICAMP), Av. Cândido Rondon, 501, Campinas 13083-875, SP, Brazil
*
Authors to whom correspondence should be addressed.
Agronomy 2025, 15(7), 1741; https://doi.org/10.3390/agronomy15071741 (registering DOI)
Submission received: 14 June 2025 / Revised: 8 July 2025 / Accepted: 16 July 2025 / Published: 19 July 2025
(This article belongs to the Special Issue Old Challenges and Modern Solutions in Farmland Soils: Addressing Heavy Metals, Microplastics, and GHG Emissions)
Browse Figures
Versions Notes

Abstract

Rare earth elements (REEs) are emerging soil contaminants due to increasing fertilizer use, mining activities, and technological applications. However, few studies have assessed their concentrations in soils or associated environmental risks. Here, we evaluate the influence of land cover types (Eucalyptus plantation, forest, and pasture), parent material, and soil physicochemical properties (predictor variables) on REE content in the Brazilian Atlantic Forest and measure pseudo-total REE content using inductively coupled plasma mass spectrometry (ICP-MS). Differences in REE content across land cover types, parent materials, and soil properties were assessed using similarity and variance analyses (ANOSIM, ANOVA, and Kruskal–Wallis) followed by post hoc tests (Tukey HSD and Dunn’s). We used model selection based on the Akaike criterion (ΔAICc < 2) to determine the influence of predictor variables on REE content. Our results showed that parent materials (igneous and metamorphic rocks) were the best predictors, yielding plausible models (Adj R2 ≥ 0.3) for Y, δEu, and LaN/SaN. In contrast, Ca:Mg alone provided a plausible model (Adj R2 = 0.15) for δCe anomalies, while clay content (Adj R2 = 0.11) influenced the SaN/YbN ratio, though soil properties had weaker effects than parent materials. However, we found no evidence that Eucalyptus plantations or pastures under non-intensive management increase REE content in Brazilian Atlantic Forest soils.

1. Introduction

The conversion of natural vegetation into agricultural, livestock, mining, and industrial areas has led to severe environmental degradation, including soil and water pollution [1,2]. Unsustainable agricultural practices, in particular, have increased heavy metal concentrations in soils, threatening biodiversity [3], ecosystem health, and essential ecological services [2,4,5]. Among the primary sources of heavy metal contamination are the excessive use of fertilizers, pesticides, animal feeds, and contaminated irrigation water [1,5], which introduce metals such as cadmium, mercury, copper, zinc, nickel, and chromium into soils [6].
Globally, another growing concern is the rise in rare earth elements (REEs) as emerging environment contaminants [7,8]. The REE group consists of 14 lanthanides, classified as light (La, Ce, Pr, Nd, Sm, and Eu) and heavy (Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu), along with scandium (Sc) and yttrium (Y) [9,10,11]. Modern industry is primarily responsible for the increased exploitation of REEs and the consequent dispersion of their concentrations in the environment [12]. Furthermore, higher contents of REEs in soils are also associated with the phosphate fertilizers use, which are from phosphate rocks and persist as impurities even after processing [13,14]. Naturally, REEs originate from parent material weathering [15], but their retention in soils is influenced by properties such as pH, organic matter, clay content, cation exchange capacity (CEC), and Fe oxides [9,16]. For instance, Fe-rich and clayey and alkaline soils with high CEC tend to retain more REEs [14,16].
Fertilizers are essential for crop productivity [1,13], but their overuse can lead to bioaccumulation in the food chain [4,17], posing risks to ecosystems, biodiversity, and human health [2,3]. In extreme cases, soils may become toxic, losing their capacity to sustain plant growth. Elevated REEs’ levels can lead to bioaccumulation in plants, potentially contaminating the food chain and posing ecological and health risks [8,10]. In Brazil, REE accumulation is primarily linked to mining, phosphate fertilizer use, and effluent discharge [8,15,18]. Brazilian soils, typically low in fertility, require heavy phosphate fertilizer application to maintain crop yields [14,19]. However, prolonged use of these fertilizers may elevate REE concentrations [17,18].
Nowadays, the main concern about REEs spreading is the lack of information for safe guidelines for REE thresholds and about their long-term effects on wildlife, soil [18] and water quality, and human health due to bioaccumulation in the environment [20]. Studies suggest that cultivated soils often contain higher REE levels than natural vegetation areas [16]. In Brazil, studies to predict the REE distribution across land use types remain limited [16]. Most studies have focused on describing the relationships between REEs and parent materials [10,11] or characterizing mining impacts [8], often neglecting the role of land management in fertilizer-driven REE accumulation [16] in different production systems. This knowledge gap hinders effective monitoring of soil health and biodiversity in agricultural expansion zones [18]. In this context, this study evaluates how soils in different land cover types (Eucalyptus, native forest, and pasture), parent materials, and soil properties influence REE concentrations in Brazil’s Atlantic Forest, a biodiversity hotspot [21]. Although the region has a long history of human occupation, current land uses are less intensive, potentially minimizing anthropogenic REE contributions. We address the following questions:
  • Do Eucalyptus and pasture in long-established agricultural lands increase REE concentrations in southeastern Brazilian soils?
  • Which predictors (land cover, parent material, soil properties) best explain variations in soil REE content?
  • What is the relative influence of these predictors on REE distribution?

2. Materials and Methods

2.1. Study Area

The study was conducted in southeastern Brazil, encompassing the northwestern region of São Paulo state within the Atlantic Forest ecoregion (Figure 1). Our sampling sites were located within the long-term ecological research project Corredor Cantareira (PELD-CCM) [22], which covers 727,000 hectares across 42 municipalities in São Paulo state. The landscape is characterized by natural forest vegetation (35%), combined agriculture/pasture systems (25%), pasture (18%), urban areas (12%), Eucalyptus and other non-native tree plantations (6%), temporary crops (2%), water bodies (1.5%), and other minor land covers (0.5%).
The region experiences a Cwa climate (humid subtropical) under the Köppen classification, featuring dry winters and hot summers with mean annual precipitation of 1800 mm. Elevation across the study area ranges from 760 to 1235 m above sea level. We used the Brazilian geological map at 1:1,000,000 scale to identify the parent materials in the study area [23] (Figure 2). The Varginha-Guaxupé unit was further subdivided due to distinct lithological characteristics. Notably, sampling sites in Granito Piracaia extend slightly into neighboring Minas Gerais state.

2.2. Soil Sampling and Analysis of Structural and Physicochemical Properties

We collected 144 soil samples (0–20 cm depth) from 48 sites across different land cover types between November 2023 (6 sites) and May (26 sites)–July 2024 (16 sites) for pseudo-total REE content analysis. In total, we assessed 48 soil samples derived from two major parent material categories: igneous (N = 25) and metamorphic (N = 23) rocks (Figure 2). The sampling design included balanced representation across land cover types, with 12 sites in Eucalyptus plantations, 24 in native forests, and 12 in pasture areas (Figure 1).
Composite (disturbed) soil samples from each site were used to determine chemical properties including pH (0.01 mol·L−1 CaCl2), exchangeable Al, Ca, and Mg (1 mol·L−1 KCl), available K and P (Mehlich-1 extractant), and organic carbon content (Walkley-Black colorimetric method) [22]. All chemical analyses were conducted at the Unithal laboratory in Campinas, Brazil (www.unithal.com.br).
For physical property characterization, we collected three undisturbed soil cores per site using metal rings (5 cm height × 8 cm diameter, 250 cm3 volume). Hydraulic conductivity was measured using a KSAT instrument (UMS GmbH, Munich, Germany) after 24 h saturation with distilled water. Bulk density was determined by oven-drying samples at 105 °C for 48 h [24]. Particle size distribution was analyzed using the pipette method (clay and silt fractions) and sieving (sand fractions). Particle density was measured using volumetric flasks, and microporosity was assessed via tension table [25].
Soil penetration resistance was measured with an electronic penetrometer (MA 933, MARCONI®) after samples reached equilibrium at field capacity (−10 kPa) [26]. All physical analyses were performed at the Laboratório de Física do Solo, University of Campinas (www.feagri.unicamp.br).

2.3. Determination of Rare Earth Elements (REEs)

Pseudo-total REE content in soils was determined using an inductively coupled plasma mass spectrometer (Thermo Scientific™ iCAP Q ICP-MS, Germany, Bremen). A total of 144 soil samples (72 from forests, 36 from pastures, and 36 from Eucalyptus plantations) representing 48 sites (three samples per site) were microwave-digested following EPA Method 3051A [27]. Site-level REE concentrations were calculated as the mean of three subsamples.
For digestion, each 0.2 g soil sample was treated with 7.5 mL concentrated nitric acid (65% v/v, ultrapure) and 2.5 mL concentrated hydrochloric acid (37% v/v, ultrapure). The resulting solution was diluted tenfold with Milli-Q water (18.2 MΩ·cm) to obtain a 2% v/v acid matrix (75% HNO3/25% HCl) prior to ICP-MS analysis. Quality control measures included the following: (1) internal standardization using 100 μg/L of 115In and 209Bi; (2) external calibration with a five-point analytical curve (0.1–20 mg·kg−1) prepared from a multi-element standard (Inorganic Ventures, 100 mg·kg−1 REEs in 2% HNO3); and (3) daily instrument optimization using Tune B solution (1 ng·mL−1 Li, Co, In, Ba, Ce, Bi, U in 2% HNO3/0.5% HCl; Thermo Scientific, Germany, Bremen).
Method accuracy was verified using IAG UoK Loess reference material (v1.00, February 2017), which showed excellent recovery for light REEs (89–94%) and progressively lower recovery for heavy REEs (47–100%): 139La (89%), 140Ce (93%), 141Pr (90%), 146Nd (90%), 147Sm (94%), 153Eu (94%), 157Gd (100%), 159Tb (88%), 163Dy (70%), 165Ho (77%), 166Er (57%), 169Tm (59%), 172Yb (47%), and 175Lu (81%). Yttrium recovery was 53%, while scandium was not analyzed due to lack of reference data. Method detection limits (10× standard deviation of blanks, n = 12) ranged from 0.002 mg·kg−1 (Lu) to 0.850 mg·kg−1 (Y). All analyses were conducted at the Environmental Studies Center of São Paulo State University (CEA, UNESP, Rio Claro, Brazil).

2.4. Normalized Patterns and Fractionation of REEs, Including Ce and Eu Anomalies

We evaluated light and heavy REE fractions along with cerium (Ce) and europium (Eu) anomalies (δ) by normalizing our data to the Post-Archean Australian Shale (PAAS) reference values [28,29], which represent natural REE concentrations following the Oddo-Harkins rule. Normalized concentrations (N) of light REEs (LREE) and heavy REEs (HREE) were used to calculate the LREEN/HREEN fractionation ratio. Additionally, we determined the normalized ratios LaN/YbN, LaN/SmN, and SmN/YbN to assess relative enrichment or depletion of individual REEs in soils, where values higher than 1 indicate enrichment and values lower than 1 indicate depletion relative to PAAS. Ce and Eu anomalies were calculated according to Bau and Dulski (1996) [30] using the equations: δCe = CeN/(0.5LaN + 0.5PrN) and δEu = EuN/(SmN × DyN)/2, respectively. Following established conventions [31,32,33], δCe and δEu values below 0.8 indicate negative anomalies, while values above 1.2 indicate positive anomalies.

2.5. Statistical Analysis

We conducted a preliminary correlation analysis between response and predictor variables using the ggpairs function in R [34] to reduce multicollinearity. Variables with correlation coefficients (R) higher than 0.70 were excluded from subsequent analyses. This analysis was performed separately for physicochemical soil properties (Table 1). For REE variables, we selected the least correlated elements (Supplementary Materials, Figures S1–S6), including raw concentrations, normalized values (PAAS-normalized), and anomaly values (δ).
Differences in soil properties among land cover types were assessed using one-way ANOVA (Tukey’s HSD post hoc) for normally distributed data or the Kruskal–Wallis test (Dunn’s post hoc) for non-normal distributions. Normality and homogeneity of variance were verified using Shapiro–Wilk and Levene’s tests, where p-values lower than 0.05 indicated a non-normal distribution and non-homogeneity in data.
We employed ANOSIM [35] to evaluate differences in REE composition across land cover and geological types, using Euclidean distance matrices. The ANOSIM R statistic ranges from −1 to 1, where values approaching 1 indicate strong dissimilarity between groups (between-group differences exceed within-group differences), while values near 0 suggest minimal separation between groups (similar rank distribution within and between groups). The accompanying p-value indicates the statistical significance of the observed R value.
For modeling, we first developed full linear models for structural and physicochemical variables separately. Significant predictors (p < 0.05) were retained for parsimonious model selection using Akaike’s Information Criterion corrected for small samples (AICc). Models with ΔAICc lower than 2 were considered the most parsimonious [36]. Model weights quantified relative explanatory power.
Finally, we developed simple linear models incorporating the retained soil variables along with land cover and geological type variables. For each significant structural and physicochemical soil property identified in previous steps, we constructed the following: (1) a basic model with the individual soil variable and (2) an enhanced model incorporating land cover type as an interaction term. All categorical predictors were explicitly specified as factors to identify the most parsimonious explanatory models. All statistical procedures were implemented using R (version 4.3.1) [34].

3. Results

3.1. REE Distribution Across Land Cover Types and Parent Materials

The igneous formations (CANT, SB, SBB, ATI, PB, and MP geological units) primarily consist of basic igneous rocks, including granites, monzodiorites, and granodiorites. The metamorphic units (VG and SR) comprise metasedimentary rocks (SR), gneisses, and schists (VG), with additional meta-volcanic and metasedimentary components in the SI unit.
Overall, we found similar average concentrations of rare earth elements (REEs) in soils across different land cover types. Likewise, the ANOSIM test confirmed no significant differences in REE content between all land cover types, with all R values being negative or near zero and p-values < 0.05 (Table 2 and Table 3). Positive anomalies were observed for δCe and δEu, with average values exceeding 1.2 across all land cover types (Table 2). Additionally, Ce was more abundant than other REEs in all soil types. ∑LREE consistently dominated over heavy ∑HREE (Table 2). The mean ∑REE concentrations were 186.88 mg·kg−1 (Eucalyptus plantation), 182.12 mg·kg−1 (forest), and 202.50 mg·kg−1 (pasture)—higher than the reference value for acid igneous rocks (102.8 mg·kg−1) but lower than that for metamorphic rocks (274 mg·kg−1) [16].
Similarly, mean ∑LREE values (177.84 mg·kg−1 for Eucalyptus, 174.07 mg·kg−1 for forest, and 193.08 mg·kg−1 for pasture) surpassed the acid igneous rock reference (92 mg·kg−1) but remained below the metamorphic rock reference (246 mg·kg−1). The same trend held for ∑HREE, with mean values of 9.03 mg·kg−1 (Eucalyptus), 8.05 mg·kg−1 (forest), and 9.42 mg·kg−1 (pasture), all above the acid igneous reference (6.6 mg·kg−1) but below the metamorphic reference (20.5 mg·kg−1) [16].
In contrast to land cover types, we observed significant dissimilarities in REE concentrations between different parent material types. The highest R values from the ANOSIM test corresponded to the LaN/SmN ratio (R = 0.31), followed by δEu (R = 0.24) and Eu (R = 0.20) (Table 3). Soil sampling distribution showed that of the 12 Eucalyptus soil samples, 6 were derived from igneous rocks and 6 from metamorphic rocks. Forest soils comprised 24 samples (13 igneous, 11 metamorphic), while pasture soils included 12 samples equally divided between igneous and metamorphic parent materials (Figure 3).
The majority of sampling sites (N = 18) were located in the metamorphic Varginha-Guaxupé (VG) geological unit. Other significant units included the igneous Granito Cantareira (CANT; N = 9) and Socorro-Bragança (SB; N = 6) units, followed by the metamorphic São Roque (SR; N = 4) unit. Additional sampling sites were distributed across igneous units, including Granitóide Serra Barro Branco (SBB), Granito Atibaia (ATI; N = 3 each), Granito Pedra Branca (PB), and Monzodiorito Piracaia (MP; N = 2 each), with a single sample from the metamorphic Serra do Itaberaba (SI) unit.
Figure 3 presents the Post-Archean Australian Shale (PAAS)-normalized REE patterns across different land cover types. Normalized values higher than 1 indicate enrichment relative to PAAS, while values lower than 1 indicate depletion. Cerium showed the highest enrichment across all land covers (CeSoil/CePAAS = 1.19 in Eucalyptus, 1.17 in forest, and 1.35 in pasture). Pasture soils additionally exhibited lanthanum enrichment (LaSoil/LaPAAS = 1.10) and near-unity gadolinium values (GdSoil/GdPAAS = 0.97). Forest soils showed near-unity lanthanum values (LaSoil/LaPAAS = 1.03). All other REEs displayed normalized values lower than 1, indicating slight depletion, with HREEs (excluding Gd) showing values lower than 0.6, reflecting strong depletion in the soils under investigation (Figure 3).
The observed Ce enrichment relative to PAAS (Figure 3) appears unrelated to land use but rather reflects inherited characteristics from parent materials, as multiple geological units exhibited Ce enrichment compared to PAAS (Supplementary Materials—Figure S7). The Ca:Mg versus δCe model (Supplementary Materials—Figure S8A) suggests a potential relationship between increased Ce anomalies and soils derived from Ca-dominated rather than Mg-dominated rocks.
PAAS-normalized REE patterns revealed significant fractionation between ETRPs and ETRLs across all geological units (Supplementary Materials—Figure S7). The PB unit showed particularly strong Ce enrichment (CeSoil/CePAAS = 2.77), while MP exhibited La enrichment (LaSoil/LaPAAS = 1.9). Additional Ce enrichment occurred in ATI and MP units (CeSoil/CePAAS = 1.78 and 1.73, respectively). Europium patterns varied substantially, with enrichment in CANT (EuN > 1.2), near-unity values in PB and MP, and pronounced depletion in metamorphic SR and SI units.

3.2. Influence of Land Cover Types, Parent Materials, and Soil Properties on REE Content

Parent materials exerted a strong influence on REE content, particularly for yttrium (Figure 4A), δEu (Figure 4B), and LaN/SmN ratios (Figure 4C). Igneous rocks from the Monzodiorito Piracaia (MP; estimate = 11.79) and Granito Pedra Branca (PB; estimate = 15.50) units showed significant positive effects on yttrium content (Model adj R2 = 0.37). For δEu, positive influences were observed from igneous Granitóide Serra Barro Branco (SBB; estimate = 2.89) and metamorphic Serra do Itaberaba (SI; estimate = 2.84) and São Roque (SR; estimate = 1.93) units (model adj R2 = 0.30).
Conversely, significant negative effects on LaN/SmN ratios were found for igneous units (CANT: estimate = −0.65; SB: −0.41; SBB: −0.40) and metamorphic units (SI: −0.64; SR: −0.53; VG: −0.61) (model adj R2 = 0.38; Table 4; Figure 4). Soil properties showed weaker associations, with the Ca:Mg ratio positively influencing Δ Ce (estimate = 0.29; adj R2 = 0.15) and clay content negatively affecting SmN/YbN ratios (estimate = −0.01; adj R2 = 0.11) (Table 4; Figure S8A,B).
Phosphorus content significantly increased ∑HREE levels (estimate = 3.17; adj R2 = 0.26), with land cover interactions revealing distinct patterns. Eucalyptus soils showed enhanced ∑HREE content with higher phosphorus, while forest (estimate = −3.00) and pasture (estimate = −2.67) soils exhibited neutral to weakly negative responses (Table 4; Figure S8C).

3.3. Soil Properties Across Land Cover Types

The analysis revealed more pronounced differences in physical soil properties (bulk density, macroporosity, total porosity, and penetration resistance) than in chemical characteristics (base saturation and Ca:Mg ratio) among the three land cover types. Pasture soils showed the most distinct properties, exhibiting significantly higher bulk density (p = 0.0031; Figure 5A), penetration resistance (p < 0.0001; Figure 5B), base saturation (p = 0.0267; Figure 5C), and Ca:Mg ratio (p = 0.0452; Figure 5D) compared to forest soils, while demonstrating lower macroporosity (p = 0.0070; Figure 5E). Similarly, pastures displayed greater penetration resistance (p = 0.0038), base saturation (BS) (p = 0.0054), and Ca:Mg ratio (p = 0.0172) relative to Eucalyptus plantations (Figure 5B–D). Forest and Eucalyptus soils differed only marginally in total porosity (p = 0.0564; Figure 5F), with forest soils showing slightly higher values.

4. Discussion

Our analysis of 144 subsamples from 48 sites across Eucalyptus, forest, and pasture soils revealed that parent material—particularly acidic igneous (N = 25) and metamorphic rocks (N = 23)—strongly influenced REE content, while land cover showed no significant effect. Soil properties, including phosphorus, clay content, and Ca:Mg ratio showed weak predictive power for REE content. Contrary to findings by Bispo et al. (2021) [16], who reported significant REE differences between productive and natural soils, our results align with Paye et al. (2016) [10] and Mihajlovic & Rinklebe (2018) [37] in identifying parent material and soil properties as primary REE determinants.
Our analysis revealed particularly strong geochemical signatures for cerium (Ce) and europium (Eu) across all land covers (δCe > 1.2; δEu > 1.3), consistent with their known abundance in Earth’s crust [9,15], along with light REEs collectively [14] and yttrium (Y) [16]. Pasture and Eucalyptus soils contained higher total REE concentrations than forests, with pasture soils particularly enriched in lanthanum (La). Light REEs consistently dominated over heavy REEs, with slight gadolinium (Gd) enrichment observed at some sites (Figure 3). These patterns reflect Brazil’s diverse geological background and soil types [19], confirming the predominant control of lithogenic factors over land use in REE distribution.
In intensive farming systems, the highest concentrations of Ce and other REEs are often linked to the excessive use of phosphate fertilizers. In our study area, land use is characterized by a mosaic of non-intensive farming systems alongside remaining forest patches. Previous research has shown that non-intensive pasture systems in this region support bird diversity [38], enhance soil microbial activity [39], and maintain genetic variability in the pioneer tree Cecropia hololeuca due to high seed dispersal permeability [38]. Similarly, most Eucalyptus plantations in the area are managed under low-intensity systems, such as the coppice systems observed in this study [40,41].
We found that parent materials had a higher influence on REE content in the study area. This pattern is related to the predominance of low-managed farming systems, which receive lower input of p-fertilizers, revealing that the type of farming system may play a key role in explaining REE concentrations. In this context, further research is needed to assess whether REEs and related proxies can serve as reliable soil quality indicators, particularly in reflecting the farming intensity history of a given area. Evaluating the impact of different agricultural systems and management practices on REE content will be essential for this purpose.
The most plausible models (ΔAICc < 2) revealed a strong influence of parent material composition on yttrium (Figure 4A), δEu (Figure 4B), and LaN/SmN (Figure 4C; Table 4). These findings align with the known geochemical association between specific REEs and particular rock formations. Additionally, we identified a weak positive correlation between the Ca:Mg ratio and δCe (Figure S8A) and a weak negative relationship between clay content and the SmN/YbN (Figure S8B; Table 4).
The inverse relationship between clay content and the SmN/YbN ratio suggests preferential enrichment of heavy HREEs in clay-rich soil profiles (Figure S8B). This pattern occurs because HREEs are more readily adsorbed than LREEs, leading to the formation of ion-adsorbed clays that concentrate HREEs [42]. Such ion-adsorbed clays primarily form through weathering of granitic and other volcanic rocks [42]. In contrast, Ramos et al. (2016) [15] demonstrated that HREEs tend to associate with weathering-resistant minerals, explaining their persistence in sandy soils, while LREEs show greater affinity for clay-rich soils [9]. Mihajlovic & Rinklebe (2018) [37] also found a significant relationship between ∑REE and soil clay content. However, the influence of soil clay on the concentrations and fractionation of REEs has yet to be fully elucidated.
Notably, Y enrichment in the MP and PB geological units reflects their igneous rock provenance, which typically contains Y-rich minerals [43]. Similarly, we observed LREE enrichment (LaN/SmN > 1.5) in the ATI, PB, and MP units, all derived from acidic igneous sources. In granitic systems, accessory minerals like monazite [(Ce,La)PO4] play a crucial role in LREE enrichment. These minerals typically contain La (~23.79%), Ce (~42.33%), Pr (~3.01%), Nd (~3.45%), and Gd (~0.66%), with the remaining composition consisting of P, U, Th, and Si [43].
We also observed a significant positive correlation between phosphorus content and heavy ∑HREE concentrations in Eucalyptus plantation soils. This relationship appears strongly influenced by an outlier sample showing exceptionally high ∑HREE values (Figure S8C). The mechanisms underlying the observed associations with both the Ca:Mg ratio and phosphorus content remain unclear due to limited comparable studies in the literature. Regarding Ce anomalies and the Ca:Mg ratio, we propose the following two potential explanations: (1) a geogenic origin linked to parent material composition or (2) an anthropogenic effect resulting from intensive calcium fertilizer application (Figure 5D). While Moreira et al. (2019) [19] reported minimal ecotoxicological risks from Ce in Cerrado ecosystems due to natural soil properties that limit its bioavailability, our findings suggest potential environmental concerns in areas with naturally elevated REEs. The combination of geological REE enrichment and increasing phosphate fertilizer use may create cumulative environmental and human health risks that warrant monitoring. Of particular concern is the conversion of native forests to intensive agricultural systems, which typically require greater phosphate inputs and may consequently elevate REE concentrations. These findings highlight the urgent need for environmental policies regulating REE levels and establishing monitoring protocols for affected regions.
Regarding soil properties, our analysis revealed the most significant differences between pasture–forest and pasture–Eucalyptus soil comparisons. Pasture soils exhibited higher density (Figure 5A), penetration resistance (Figure 5D), BS (Figure 5E), and Ca:Mg ratio (Figure 5F) compared to forest soils. Similarly, pasture soils showed greater penetration resistance (Figure 5D), BS (Figure 5E), and Ca:Mg ratio (Figure 5F) than Eucalyptus soils. Conversely, forest soils maintained superior macroporosity (Figure 5C) and total porosity relative to both pasture and Eucalyptus soils (Figure 5B). These findings align with established knowledge that non-sustainable agricultural practices significantly alter soil structural and physicochemical properties [44].
The observed patterns in Eucalyptus soils—particularly lower BS, reduced Ca and Mg levels (Figure 5E,F), and modified organic matter fractions [45] compared to forest and pasture soils—are consistent with previous research and help explain our results. In the Brazilian context, pasture-associated soil degradation has been well-documented, primarily due to low stocking rates combined with overgrazing [46]. These practices, along with intensive grazing pressure and cattle trampling [46,47], lead to increased soil density and compaction. Additionally, machinery use represents another significant factor contributing to elevated soil density and reduced porosity [48].
Forest soils demonstrate contrasting characteristics, exhibiting higher porosity (Figure 5B,C) and lower penetration resistance (Figure 5D), attributable to greater root biomass, enhanced organic carbon content, and reduced compaction [48]. Notably, Leite et al. (2010) [45] observed that soils under long-term Eucalyptus plantations show gradual recovery of organic matter content compared to pasture systems, though our study found no statistically significant differences in carbon stocks across land uses. Instead, carbon stocks showed strong correlations with clay content and phosphorus levels (Figure S2). These patterns likely reflect management practices in our study area—Eucalyptus stands predominantly managed through coppice systems, and pastures maintained as mixed grass communities dominated by Brachiaria spp. with secondary components like Andropogon bicornis.
We found differences in soil quality across different land covers analyzed. Thus, our findings revealed that although the Eucalyptus and pasture seem like non-intensive systems, financial incentives are necessary to guide the establishment of sustainable production systems and management practices to help establish soil health in the study area.
Notably, we found no evidence that non-intensive Eucalyptus or pasture systems increase soil REE concentrations in our study area. Our data reveal that in this part of the Atlantic Forest biome, parent material consistently determines REE concentrations across both agricultural and non-agricultural soils under the same weathering regime.
Our findings suggest that the capacity of land use to modify soil REE patterns may have been overestimated by Bispo et al. (2021) [16], who attributed agricultural soil REE content primarily to land use while ascribing non-agricultural soil REEs to parent materials. Although REE content in agricultural soils are mainly defined because parent material, increases in REEs concentrations in soils due to agricultural activity depends on the production systems, highlighting that additional studies in other agricultural systems need to be performed in Brazil to help establish safe REEs thresholds.

5. Conclusions

Our findings demonstrate that parent materials and physicochemical soil properties are the most significant predictors of rare earth element (REE) concentrations in the study area, with geological factors exerting a stronger influence than soil characteristics. This strong geogenic control suggests that natural REE background levels could serve as valuable baselines for detecting anthropogenic REE inputs and informing environmental policy development to establish safe thresholds for these elements in ecosystems. Importantly, although our study area is dominated by non-intensive farming systems, we still observed measurable impacts of pasture management on soil quality indicators.
While our results generally align with previous reports of REE abundance patterns, we note substantial variability in published REE concentrations across different Brazilian regions. This heterogeneity complicates the identification of clear spatial patterns and makes meaningful comparisons of REE levels particularly challenging.
Undoubtedly, further research is needed to establish natural background levels of REEs and inform environmental policy development for regions with varying geological REE concentrations. We specifically recommend that future studies include the following: (1) encompassing diverse land cover types and (2) thoroughly documenting production system types and management practices. Such approaches will enable the generation of comparable REE concentration baselines across Brazil’s heterogeneous territories.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/agronomy15071741/s1, Figure S1: Influence of soil properties, parent material, and land cover types on Rare Earth Elements within Atlantic Forest, Brazil; Figure S2: Correlation matrix between physical-chemical soil variables with the coefficient of correlation less than 0.7; Figure S3: Correlation matrix between physical soil variables; Figure S4: Correlation matrix between physical soil variables with the coefficient of correlation less than 0.7; Figure S5: Correlation matrix between the rare earth elements (response variables); Figure S6: Correlation matrix between the rare earth elements (response variables) with the coefficient of correlation less than 0.7; Figure S7: Enrichment and depletion of REEs in the different parent materials. The semi-total values of REEs were normalized according to Post-Archean Australian Shale (PAAS); Figure S8: Relationship between rare earth elements and structural and physical-chemical soil variables. The results are referent to plausible models selected from model selection, using the AIC criterion for Ca:Mg, clay, and phosphorus contents; Table S1: Results of Levenes, Shapiro-wilk, ANOVA, Kruskal-Wallis, and post hoc tests used to assess the significant differences between structural and physical-chemical soil variables in Eucalyptus, forest, and pasture; Table S2: Results of model selection based on AIC criterion considering rare earth elements and derived indexes as response variables and structural and physical-chemical soil variables as predictors variables. In gray are highlighted the significant variables (p-value < 0.05) that were used to compose the final models with land covers and geological types variables; Table S3: List of final models, including plausible and non-plausible models according to model selection based on AIC criterion. The final plausible models are also shown in the main body of the manuscript.

Author Contributions

Conceptualization, G.R.C., J.S.d.S. and A.A.M.; methodology, G.R.C., J.S.d.S., M.B.T.Z., R.S.C.A., Z.M.d.S., M.C.R. and A.A.M.; software, J.S.d.S.; validation, G.R.C., J.S.d.S. and M.B.T.Z.; formal analysis, G.R.C., J.S.d.S., M.B.T.Z. and Z.M.d.S.; investigation, G.R.C., J.S.d.S., M.B.T.Z., R.S.C.A., Z.M.d.S., M.C.R. and A.A.M.; resources, M.C.R. and A.A.M.; data curation, G.R.C., J.S.d.S., M.B.T.Z. and Z.M.d.S.; writing—original draft preparation, G.R.C. and J.S.d.S.; writing—review and editing, M.B.T.Z., R.S.C.A., Z.M.d.S., M.C.R. and A.A.M.; visualization, G.R.C., J.S.d.S., M.B.T.Z., R.S.C.A., Z.M.d.S., M.C.R. and A.A.M.; supervision, M.C.R. and A.A.M.; project administration, M.C.R. and A.A.M.; funding acquisition, M.C.R. and A.A.M. All authors have read and agreed to the published version of the manuscript.

Funding

G.R.C. received financial support through a CAPES doctoral scholarship. J.S.S received financial support through a FAPESP postdoctoral scholarship (process #2019/09713-6), a CAPES-Print scholarship (process #88887.890889/2023-00), and a postdoctoral grant from the National Council for Scientific and Technological Development (CNPq, process #179354/2024-8). MCR expresses sincere gratitude to the São Paulo Research Foundation (FAPESP processes #2013/50421-2; #2020/01779-5; #2021/06668-0; #2021/08322-3; #2021/08534-0; #2021/10195-0; #2021/10639-5; #2022/10760-1) and the National Council for Scientific and Technological Development (CNPq processes #442147/2020-1; #402765/2021-4; #313016/2021-6; #440145/2022-8; #420094/2023-7; #446029/2024-6; #421464/2025-9), as well as São Paulo State University (UNESP) for their substantial financial support.

Data Availability Statement

Data available upon request.

Acknowledgments

We gratefully acknowledge the Sistema Integrado de Gestão Ambiental do Estado de São Paulo (SIGAM) and Secretaria de Meio Ambiente, Infraestrutura e Logística do Estado de São Paulo (SEMIL) for approving the fieldwork conducted within protected areas. We would also like to thank Landerlei Almeida Santos and Luiz Roberto Camargo Ribeiro for their support in preparing the manuscript. This research benefited from the valuable infrastructure and support provided by the Pierre Kaufmann Radio Observatory, located in Atibaia, São Paulo, which is maintained and operated by the Mackenzie Presbyterian University through its Mackenzie Radio Astronomy and Astrophysics Center (CRAAM) in collaboration with the National Institute for Space Research (INPE). We extend our special appreciation to Guilherme Alaia for his ongoing support to the LTER CCM team. This study was conducted as part of the Center for Research on Biodiversity Dynamics and Climate Change, which receives funding from the São Paulo Research Foundation (FAPESP).

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Rashid, A.; Schutte, B.J.; Ulery, A.; Deyholos, M.K.; Sanogo, S.; Lehnhoff, E.A.; Beck, L. Heavy Metal Contamination in Agricultural Soil: Environmental Pollutants Affecting Crop Health. Agronomy 2023, 13, 1521. [Google Scholar] [CrossRef]
  2. Luo, H.; Yang, L.; Zhang, C.; Xiao, X.; Lyu, X. Early warning of heavy metals contamination in agricultural soils: Spatio-temporal distribution and future trends in the Hexi Corridor. Ecol. Indic. 2024, 160, 111908. [Google Scholar] [CrossRef]
  3. Barbosa, M.D.; Fernandes, A.C.; Alves, R.S.; Alves, D.A.; Barbosa Júnior, F.; Batista, B.L.; Ribeiro, M.C.; Hornos Carneiro, M.F. Effects of native forest and human-modified land covers on the accumulation of toxic metals and metalloids in the tropical bee Tetragonisca angustula. Ecotoxicol. Environ. Saf. 2021, 215, 112147. [Google Scholar] [CrossRef] [PubMed]
  4. Wuana, R.A.; Okieimen, F.E. Heavy Metals in Contaminated Soils: A Review of Sources, Chemistry, Risks and Best Available Strategies for Remediation. ISRN Ecol. 2011, 2011, 1–20. [Google Scholar] [CrossRef]
  5. Angon, P.B.; Islam, M.S.; Kc, S.; Das, A.; Anjum, N.; Poudel, A.; Suchi, S.A. Sources, effects and present perspectives of heavy metals contamination: Soil, plants and human food chain. Heliyon 2024, 10, e28357. [Google Scholar] [CrossRef] [PubMed]
  6. Nde, S.C.; Félicité, O.M.; Aruwajoye, G.S.; Palamuleni, L.G. A meta-analysis and experimental survey of heavy metals pollution in agricultural soils. J. Trace Elem. Miner. 2024, 9, 100180. [Google Scholar] [CrossRef]
  7. Borowiak, K.; Lisiak, M.; Kanclerz, J.; Budka, A.; Mleczek, M.; Niedzielski, P.; Adamska, A.; Janicka, E. Relations between rare earth elements accumulation in Taraxacum officinale L. and land use in an urban area—A preliminary study. Ecol. Indic. 2018, 94, 22–27. [Google Scholar] [CrossRef]
  8. Pereira, W.V.; Ramos, S.J.; Melo, L.C.; Braz, A.M.; Dias, Y.N.; Almeida, G.V.; Fernandes, A.R. Levels and environmental risks of rare earth elements in a gold mining area in the Amazon. Environ. Res. 2022, 2011, 113090. [Google Scholar] [CrossRef] [PubMed]
  9. Silva, F.B.; Nascimento, C.W.; Álvarez, A.M.; Araújo, P.R. Inputs of rare earth elements in Brazilian agricultural soils via P-containing fertilizers and soil correctives. J. Environ. Manag. 2019, 232, 90–96. [Google Scholar] [CrossRef] [PubMed]
  10. Paye, H.D.; Mello, J.W.; Mello, J.W.; Mascarenhas, G.M.; Gasparon, M.; Gasparon, M. Distribution and fractionation of the rare earth elements in Brazilian soils. J. Geochem. Explor. 2016, 161, 27–41. [Google Scholar] [CrossRef]
  11. Brilhante, S.A.; da Silva, Y.J.; de Medeiros, P.L.; do Nascimento, C.W.; da Silva, Y.J.; Ferreira, T.O.; Otero, X.L.; Silva, A.H.; Sousa, M.G.; Alcântara, V.; et al. Geochemistry of rare Earth elements in rocks and soils along a Cretaceous volcano-sedimentary Basin in Northeastern Brazil. Geoderma Reg. 2024, 26, e00756. [Google Scholar] [CrossRef]
  12. Martina, A.; Ferroni, L.; Marrocchino, E. The Soil–Plant Continuity of Rare Earth Elements: Their Interactions with Plant Structures and Processes. J. Xenobiot. 2025, 15, 46. [Google Scholar] [CrossRef] [PubMed]
  13. Silva, Y.J.; Nascimento, C.W.; Silva, Y.J.; Biondi, C.M.; Silva, C.M. Rare Earth Element Concentrations in Brazilian Benchmark Soils. Rev. Bras. Cienc. Solo 2016, 40, e0150413. [Google Scholar] [CrossRef]
  14. Ribeiro, P.G.; Dinali, G.S.; Boldrin, P.; de Carvalho, T.S.; de Oliveira, C.; Ramos, S.J.; Siqueira, J.O.; Moreira, C.G.; Guilherme, L.R. Rare Earth Elements (REEs) Rich-Phosphate Fertilizers Used in Brazil are More Effective in Increasing Legume Crops Yield Than Their REEs-Poor Counterparts. Int. J. Plant Prod. 2021, 15, 1–11. [Google Scholar] [CrossRef]
  15. Ramos, S.J.; Dinali, G.S.; Oliveira, C.D.; Martins, G.C.; Moreira, C.G.; Siqueira, J.O.; Guilherme, L.R. Rare Earth Elements in the Soil Environment. Curr. Pollut. Rep. 2016, 2, 28–50. [Google Scholar] [CrossRef]
  16. Bispo, F.H.; de Menezes, M.D.; Fontana, A.; Sarkis, J.E.; Gonçalves, C.M.; de Carvalho, T.S.; Curi, N.; Guilherme, L.R. Rare earth elements (REEs): Geochemical patterns and contamination aspects in Brazilian benchmark soils. Environ. Pollut. 2021, 289, 117972. [Google Scholar] [CrossRef] [PubMed]
  17. Turra, C.; Fernandes, E.A.N.; Bacchi, M.A. Evaluation on rare earth elements of Brazilian agricultural supplies. J. Environ. Chem. Ecotoxicol. 2011, 3, 86–92. [Google Scholar]
  18. Silva, Y.J.; Oliveira, E.B.; Silva, Y.J.; Nascimento, C.W.; Silva, T.D.; Boechat, C.L.; Teixeira, M.P.; Barbosa, R.S.; Singh, V.P.; Sena, A.F. Quality reference values for rare earth elements in soils from one of the last agricultural frontiers in Brazil. Sci. Agri. 2021, 78, e20200069. [Google Scholar] [CrossRef]
  19. Moreira, C.G.; Carvalho, T.S.; de Oliveira, C.; Abreu, L.B.; Castro, A.C.; Ribeiro, P.G.; Bispo, F.H.; Boutin, C.; Guilherme, L.R. Ecological risk assessment of cerium for tropical agroecosystems. Chemosphere 2019, 221, 124–131. [Google Scholar] [CrossRef] [PubMed]
  20. Pereto, C.; Baudrimont, M.; Coynel, A. Global natural concentrations of Rare Earth Elements in aquatic organisms: Progress and lessons from fifty years of studies. Sci. Total. Environ. 2014, 222, 171241. [Google Scholar] [CrossRef] [PubMed]
  21. Myers, N.; Mittermeier, R.A.; Mittermeier, C.G.; Fonseca, G.A.; Kent, J. Biodiversity hotspots for conservation priorities. Nature 2000, 403, 853–858. [Google Scholar] [CrossRef] [PubMed]
  22. Montagnana, P.C.; Alves, R.S.; Garófalo, C.A.; Ribeiro, M.C. Landscape heterogeneity and forest cover shape cavity-nesting hymenopteran communities in a multi-scale perspective. Basic Appl. Ecol. 2021, 56, 239–249. [Google Scholar] [CrossRef]
  23. CPRM. Geologia do Brasil. 2004. Available online: https://metadados.snirh.gov.br/geonetwork/srv/api/records/4de151f5-6d20-4e87-b49c-c9b0ea65440c (accessed on 17 July 2024).
  24. Raij, B.V.; Andrade, J.C.; Cantarella, H.; Quaggio, J.A. Análise Química para Avaliação da Fertilidade de Solos Tropicais, 1st ed.; Instituto Agronômico: Campinas, Brasil, 2001; 285p. [Google Scholar]
  25. Teixeira, P.C.; Donagemma, G.K.; Fontana, A.; Teixeira, W.G. Manual de Métodos de Análise de Solos, 3rd ed.; Embrapa: Brasília, Brasil, 2017; 574p. [Google Scholar]
  26. Otto, R.; Silva, A.P.; Franco, H.C.; Oliveira, E.C.; Trivelin, P.C. High soil penetration resistance reduces sugarcane root system development. Soil. Tillage Res. 2011, 117, 201–210. [Google Scholar] [CrossRef]
  27. U.S. EPA. Method 3051A (SW-846): Microwave Assisted Acid Digestion of Sediments, Sludges, and Oils; Revision 1; U.S. EPA: Washington, DC, USA, 2007. [Google Scholar]
  28. McLennan, S.M. Rare Earth Elements in Sedimentary Rocks: Influence of Provenance and Sedimentary Processes. In Geochemistry and Mineralogy of Rare Earth Elements; Lipin, B.R., McKay, G.A., Eds.; De Gruyter: Berlin, Germany, 1989; pp. 169–200. [Google Scholar] [CrossRef]
  29. Pourmand, A.; Dauphas, N.; Ireland, T.J. A novel extraction chromatography and MC-ICP-MS technique for rapid analysis of REE, Sc and Y: Revising CI-chondrite and Post-Archean Australian Shale (PAAS) abundances. Chem. Geol. 2012, 291, 38–54. [Google Scholar] [CrossRef]
  30. Bau, M.; Dulski, P. Distribution of yttrium and rare-earth elements in the penge and kuruman iron-formations, transvaal supergroup, South Africa. Precambrian Res. 1996, 79, 37–55. [Google Scholar] [CrossRef]
  31. Grawunder, A.; Merten, D.; Büchel, G. Origin of middle rare earth element enrichment in acid mine drainage-impacted areas. Environ. Sci. Pollut. Res. 2014, 21, 6812–6823. [Google Scholar] [CrossRef] [PubMed]
  32. Migaszewski, Z.M.; Gałuszka, A.; Dołęgowska, S. Rare earth and trace element signatures for assessing an impact of rock mining and processing on the environment: Wiśniówka case study, south-central Poland. Environ. Sci. Pollut. Res. 2016, 23, 24943–24959. [Google Scholar] [CrossRef] [PubMed]
  33. Migaszewski, Z.M.; Gałuszka, A.; Dołęgowska, S. Extreme enrichment of arsenic and rare earth elements in acid mine drainage: Case study of Wiśniówka mining area (south-central Poland). Environ. Pollut. 2019, 244, 898–906. [Google Scholar] [CrossRef] [PubMed]
  34. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2024. [Google Scholar]
  35. Clarke, K.R.; Green, R.H. Statistical Design and Analysis For A Biological Effects Study. Mar. Ecol. Prog. Ser. 1988, 46, 213–226. [Google Scholar] [CrossRef]
  36. Zuur, A.F.; Ieno, E.N.; Walker, N.; Saveliev, A.A.; Smith, G.M. Mixed Effects Models and Extensions in Ecology with R. Statistics for Biology and Health; Springer: New York, NY, USA, 1998. [Google Scholar]
  37. Mihajlovic, J.; Rinklebe, J. Rare earth elements in German soils—A review. Chemosphere 2018, 205, 514–523. [Google Scholar] [CrossRef] [PubMed]
  38. Santos, J.; Vitorino, L.C.; Fabrega Gonçalves, R.; Corrêa Côrtes, M.; Souza Cruz Alves, R.; Cezar Ribeiro, M.; Garcia Collevatti, R. Matrix dominance and landscape resistance affect genetic variability and differentiation of an Atlantic Forest pioneer tree. Landsc. Ecol. 2022, 37, 2481–2501. [Google Scholar] [CrossRef]
  39. Vasconcelos, T.C.; Londe, V.; Zanini, A.M.; Rodrigues, R.R. Comparative recovery of soil microbial activity and invertebrate abundance and richness in abandoned and planted pastures in Southeastern Brazil. Appl. Soil Ecol. 2024, 203, 105683. [Google Scholar] [CrossRef]
  40. Suguituru, S.S.; Silva, R.R.; Souza, D.R.; Munhae, C.D.; Morini, M.S. Ant community richness and composition across a gradient from Eucalyptus plantation to secondary Atlantic Forest. Biota Neotrop. 2011, 11, 369–376. [Google Scholar] [CrossRef]
  41. Brancalion, P.H.; Amazonas, N.T.; Chazdon, R.L.; Melis, J.; Rodrigues, R.R.; Silva, C.C.; Sorrini, T.B.; Holl, K.D. Exotic eucalypts: From demonized trees to allies of tropical forest restoration? J. Appl. Ecol. 2019, 57, 56–66. [Google Scholar] [CrossRef]
  42. Ding, L.; Azimi, G. Impact of particle size and associated minerals on rare earth desorption and incorporation mechanisms in a South American ion-adsorption clay. Sci. Rep. 2024, 14, 16216. [Google Scholar] [CrossRef] [PubMed]
  43. El-Metwally, A.; Abdel-Bary, A.; El-Azib, A.; Ragab, A. Geochemistry and REE tetrad effect of the acidic volcanics and associated stream sediments, southeast G. El-Hadid area, Central Eastern Desert, Egypt. Arab. J. Geosci. 2024, 17, 142. [Google Scholar] [CrossRef]
  44. Lima, H.N.; Coêlho, J.J.; Dubeux Júnior, J.C.; Santos, E.R.; Cunha, M.V.; Mello, A.C.; Santos, M.V. Soil chemical changes and resemblances in a chronosequence rainforest-sugarcane-pastureland in the Atlantic Forest biome. CERNE 2020, 26, 444–455. [Google Scholar] [CrossRef]
  45. Leite, F.P.; Silva, I.R.; Novais, R.F.; Barros, N.F.; Neves, J.C. Alterations of soil chemical properties by eucalyptus cultivation in five regions in the Rio Doce Valley. Rev. Bras. Cienc. Solo 2010, 34, 821–831. [Google Scholar] [CrossRef]
  46. Lopes, V.S.; Cardoso, I.M.; Fernandes, O.R.; Rocha, G.C.; Simas, F.N.; Moura, W.D.; Santana, F.C.; Veloso, G.V.; Luz, J.M. The establishment of a secondary forest in a degraded pasture to improve hydraulic properties of the soil. Soil. Tillage Res. 2020, 198, 104538. [Google Scholar] [CrossRef]
  47. Batista, P.H.D.; de Almeida, G.L.P.; Pandorfi, H.; de Lima, R.P.; de Medeiros, V.W.C.; da Silva, J.L.B.; de Melo, A.A.S.; Coutinho, A.S. Assessment spatial soil physical properties under rotational cattle grazing in Northeastern Brazil. CATENA 2024, 240, 108002. [Google Scholar] [CrossRef]
  48. Benevenute, P.A.; Morais, E.G.; Souza, A.A.; Vasques, I.C.; Cardoso, D.P.; Sales, F.R.; Severiano, E.D.; Homem, B.G.; Casagrande, D.R.; Silva, B.M. Penetration resistance: An effective indicator for monitoring soil compaction in pastures. Ecol. Indic. 2020, 117, 106647. [Google Scholar] [CrossRef]
Agronomy 15 01741 g001
Figure 1. Study area location. (A) Atlantic Forest region in northwestern São Paulo state, Brazil. (B) Distribution of 48 sampling sites across different land cover types (Eucalyptus plantation, native forest, and pasture).
Figure 1. Study area location. (A) Atlantic Forest region in northwestern São Paulo state, Brazil. (B) Distribution of 48 sampling sites across different land cover types (Eucalyptus plantation, native forest, and pasture).
Agronomy 15 01741 g001
Agronomy 15 01741 g002
Figure 2. Geological units and their lithological composition within the study area. Sampling sites are indicated by black point.
Figure 2. Geological units and their lithological composition within the study area. Sampling sites are indicated by black point.
Agronomy 15 01741 g002
Agronomy 15 01741 g003
Figure 3. Rare earth element (REE) enrichment and depletion patterns across land cover types, normalized to Post-Archean Australian Shale (PAAS) reference values.
Figure 3. Rare earth element (REE) enrichment and depletion patterns across land cover types, normalized to Post-Archean Australian Shale (PAAS) reference values.
Agronomy 15 01741 g003
Agronomy 15 01741 g004
Figure 4. Effects of parent materials on different REE content (Yttrium, δEu, LaN/SaN) according to model selection based on AIC criterion. Panels (A–C) show parent material influences, with boxplots displaying median (central line), interquartile range (box), and full data range (whiskers). Geological unit abbreviations: ATI (Granito Atibaia), CANT (Granito Cantareira), MP (Monzodiorito Piracaia), PB (Granito Pedra Branca), SB (Socorro-Bragança), SBB (Granitóide Serra Barro Branco), SI (Serra do Itaberaba), SR (São Roque), and VG (Varginha-Guaxupé). ** indicates statistical significance (p < 0.05).
Figure 4. Effects of parent materials on different REE content (Yttrium, δEu, LaN/SaN) according to model selection based on AIC criterion. Panels (A–C) show parent material influences, with boxplots displaying median (central line), interquartile range (box), and full data range (whiskers). Geological unit abbreviations: ATI (Granito Atibaia), CANT (Granito Cantareira), MP (Monzodiorito Piracaia), PB (Granito Pedra Branca), SB (Socorro-Bragança), SBB (Granitóide Serra Barro Branco), SI (Serra do Itaberaba), SR (São Roque), and VG (Varginha-Guaxupé). ** indicates statistical significance (p < 0.05).
Agronomy 15 01741 g004
Agronomy 15 01741 g005
Figure 5. Statistical comparison of soil physical and chemical properties (response variables in y-axis) across soils in different land cover types. (A) significant differences between Pasture-Forest. (B) significant differences between Pasture-Forest and Pasture-Eucalyptus. (C) significant differences between Pasture-Forest and Pasture-Eucalyptus. (D) significant differences between Pasture-Forest. (B) significant differences between Pasture-Forest and Pasture-Eucalyptus. (E) significant differences between Pasture-Forest. (F) significant differences between Forest-Eucalyptus. The p-values ≤ 0.05 indicate significant differences according to ANOVA/Kruskal–Wallis tests. Measured parameters include the following: bulk density (g·cm−3), total porosity and macroporosity (m3·m−3), penetration resistance (MPa), and base saturation (%).
Figure 5. Statistical comparison of soil physical and chemical properties (response variables in y-axis) across soils in different land cover types. (A) significant differences between Pasture-Forest. (B) significant differences between Pasture-Forest and Pasture-Eucalyptus. (C) significant differences between Pasture-Forest and Pasture-Eucalyptus. (D) significant differences between Pasture-Forest. (B) significant differences between Pasture-Forest and Pasture-Eucalyptus. (E) significant differences between Pasture-Forest. (F) significant differences between Forest-Eucalyptus. The p-values ≤ 0.05 indicate significant differences according to ANOVA/Kruskal–Wallis tests. Measured parameters include the following: bulk density (g·cm−3), total porosity and macroporosity (m3·m−3), penetration resistance (MPa), and base saturation (%).
Agronomy 15 01741 g005
Table 1. Variables retained after correlation analysis (R < 0.7). Normalized values (∑LREEN/∑HREEN, δCe, δEu, SmN/YbN, LaN/YbN, and LaN/SmN) were calculated relative to PAAS (Post-Archean Australian Shale), where ∑ = sum, δ = anomalies.
Table 1. Variables retained after correlation analysis (R < 0.7). Normalized values (∑LREEN/∑HREEN, δCe, δEu, SmN/YbN, LaN/YbN, and LaN/SmN) were calculated relative to PAAS (Post-Archean Australian Shale), where ∑ = sum, δ = anomalies.
Predictor Variables (R < 7)Response Variables (R < 7)
Soil physical variables
Soil density
Total porosity
Macroporosity
Microporosity
Fine sand
Total sand
Penetration resistance
Hydraulic conductivity
Silty
Clay		Rare earth elements
Europium
Cerium
Lutetium
Yttrium
∑LREEN/∑HREEN
∑LREE
∑HREE
∑REE
δCe
δEu
LaN/YbN
LaN/SmN
SmN/YbN


Soil chemical variables
Potassium content
Phosphorus content
Carbon stock
Ca:Mg
Table 2. Average (Avg) and standard deviation (SD) of rare earth element (REE) concentrations (mg·kg−1) across different land cover types in the Atlantic Forest, Brazil. LREEs (La to Eu), HREEs (Gd to Lu) and REEs (La to Lu); δCe = CeN/(0.5LaN × 0.5PrN) and δEu = EuN/(SmN × DyN)/2.
Table 2. Average (Avg) and standard deviation (SD) of rare earth element (REE) concentrations (mg·kg−1) across different land cover types in the Atlantic Forest, Brazil. LREEs (La to Eu), HREEs (Gd to Lu) and REEs (La to Lu); δCe = CeN/(0.5LaN × 0.5PrN) and δEu = EuN/(SmN × DyN)/2.
Eucalyptus (Avg ± SD)Forest (Avg ± SD)Pasture (Avg ± SD)
Europium (mg·kg−1)0.89 ± 0.600.79 ± 0.590.72 ± 0.39
Cerium (mg·kg−1)94.98 ± 68.6793.63 ± 39.59107.40 ± 35.48
Lutetium (mg·kg−1)0.14 ± 0.120.18 ± 0.100.20 ± 0.13
Yttrium (mg·kg−1)7.60 ± 9.535.40 ± 2.977.60 ± 7.43
∑LREEN/∑HREEN3.16 ± 1.272.87 ± 0.992.73 ± 0.99
∑LREE (mg·kg−1)177.84 ± 118.87174.07 ± 72.58193.08 ± 80.81
∑HREE (mg·kg−1)9.03 ± 6.318.05 ± 3.329.42 ± 4.82
∑REE (mg·kg−1)186.88 ± 124.51182.12 ± 75.30202.50 ± 85.40
δCe1.27 ± 0.591.47 ± 0.961.97 ± 1.79
δEu1.72 ± 1.311.36 ± 0.731.64 ± 2.30
LaN/YbN7.38 ± 4.176.93 ± 4.004.68 ± 1.78
LaN/SmN1.20 ± 0.221.27 ± 0.321.23 ± 0.24
SmN/YbN6.17 ± 3.515.36 ± 3.033.81 ± 1.45
Table 3. Results of the Analysis of Similarity (ANOSIM) comparing REE concentrations among different land cover types and nine geological units in the Atlantic Forest, Brazil. Gray shading highlights R values approaching 1, indicating greater dissimilarity between the compared groups.
Table 3. Results of the Analysis of Similarity (ANOSIM) comparing REE concentrations among different land cover types and nine geological units in the Atlantic Forest, Brazil. Gray shading highlights R values approaching 1, indicating greater dissimilarity between the compared groups.
Land Cover Types
REEsRSignificance
(p-Value)
Europium−0.080.01
Cerium0.090.00
Lutetium0.020.00
Yttrium−0.020.01
LREEN/HREEN0.020.00
∑LREE0.020.00
∑HREE−0.010.00
∑REE0.020.00
δCe0.000.00
δEu0.070.00
LaN/YbN−0.010.01
LaN/SmN0.010.00
SmN/YbN0.000.00
Geological Types
REEsRSignificance
(p-Value)
Europium0.200.00
Cerium0.020.00
Lutetium0.030.00
Yttrium0.180.00
LREEN/HREEN0.150.00
∑LREE0.140.00
∑HREE0.190.00
∑REE0.140.00
δCe0.050.00
δEu0.240.00
LaN/YbN0.000.00
LaN/SaN0.310.00
SaN/YbN−0.040.01
Table 4. Selected models explaining REE concentration variations in the Atlantic Forest soils (Brazil) based on AICc model selection (ΔAICc < 2). Only models with adjusted R2 > 0.10 are shown, including degrees of freedom (df) and multiple coefficients of determination (Mult R2). Complete model results are available in Table S2, while relationships between REEs and soil physicochemical variables are illustrated in Figure S8.
Table 4. Selected models explaining REE concentration variations in the Atlantic Forest soils (Brazil) based on AICc model selection (ΔAICc < 2). Only models with adjusted R2 > 0.10 are shown, including degrees of freedom (df) and multiple coefficients of determination (Mult R2). Complete model results are available in Table S2, while relationships between REEs and soil physicochemical variables are illustrated in Figure S8.
YttriumDAICcdfWeightMult R2Adj R2
Parent material0.00100.980.480.37
∑HREEDAICcdfWeightMult R2Adj R2
Phosphorus: land cover types0.007.00.950.340.26
δCeDAICcdfWeightMult R2Adj R2
Ca:Mg0.003.00.710.170.15
δEuDAICcdfWeightMult R2Adj R2
Parent material0.00100.850.420.30
LaN/SaNDAICcdfWeightMult R2Adj R2
Parent material0.00100.990.490.38
SaN/YbNDAICcdfWeightMult R2Adj R2
Clay0.003.00.800.130.11
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Castellano, G.R.; Santos, J.S.d.; Zanatta, M.B.T.; Alves, R.S.C.; Souza, Z.M.d.; Ribeiro, M.C.; Menegário, A.A. Rare Earth Elements in Tropical Agricultural Soils: Assessing the Influence of Land Use, Parent Material, and Soil Properties. Agronomy 2025, 15, 1741. https://doi.org/10.3390/agronomy15071741

AMA Style

Castellano GR, Santos JSd, Zanatta MBT, Alves RSC, Souza ZMd, Ribeiro MC, Menegário AA. Rare Earth Elements in Tropical Agricultural Soils: Assessing the Influence of Land Use, Parent Material, and Soil Properties. Agronomy. 2025; 15(7):1741. https://doi.org/10.3390/agronomy15071741

Chicago/Turabian Style

Castellano, Gabriel Ribeiro, Juliana Silveira dos Santos, Melina Borges Teixeira Zanatta, Rafael Souza Cruz Alves, Zigomar Menezes de Souza, Milton Cesar Ribeiro, and Amauri Antonio Menegário. 2025. "Rare Earth Elements in Tropical Agricultural Soils: Assessing the Influence of Land Use, Parent Material, and Soil Properties" Agronomy 15, no. 7: 1741. https://doi.org/10.3390/agronomy15071741

APA Style

Castellano, G. R., Santos, J. S. d., Zanatta, M. B. T., Alves, R. S. C., Souza, Z. M. d., Ribeiro, M. C., & Menegário, A. A. (2025). Rare Earth Elements in Tropical Agricultural Soils: Assessing the Influence of Land Use, Parent Material, and Soil Properties. Agronomy, 15(7), 1741. https://doi.org/10.3390/agronomy15071741

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop