Introduction

Rare earth elements (REEs) are composed of lanthanides (La-Lu), scandium, and yttrium, with a total of 17 metallic elements. According to Oddo-Harkinsthe, even atomic-numbered elements have an order of magnitude greater abundance (higher Clarke concentrations) than their adjacent odd-numbered counterparts, and further emphasized by a greater number of isotopes. On the basis these properties, REEs are classified as light rare earths (LREEs, La-Lu) and heavy rare earths (HREEs, Gd and Tb-Lu)1,2,3. Because of its physical and chemical similarities, yttrium (Y) is typically classified as HREEs, although it is the lightest REE. REEs are essential strategic resources worldwide due to their unique physicochemical properties. They are extensively used in both high-tech and traditional industries, including the national defense and military, petroleum refining, permanent magnets, textiles, and others. REEs have made important contributions to the development of the national economy4,5.

Ion-adsorption type REEs (ionic REEs), the most important REEs sources in the world, are uniquely formed from weathered granite crusts containing REEs minerals in warm and humid climates6. The primary reservoirs of global ionic REEs are located in China, particularly in southern Jiangxi Province. This region possesses substantial ionic rare earth deposits7especially those rich in terbium, dysprosium, europium, yttrium, and other high-value medium and heavy REEs8. Ionic REEs can be extracted through a leaching technique with electrolyte solutions though an ion-exchange process due to their specific form9. In-situ leaching is the most effective technology to extract REEs, with a high leaching efficiency of up to 85–90%. It has become the standard mining technology for ionic REEs ores10. The most commonly used leaching solution is 3–5% ammonium sulfate ((NH4)2SO4), with the leaching process spanning 150–400 days. In recent years, the development of the high-tech REEs industry and the global use of REEs-related products have promoted large-scale exploitation of rare earth resources, and also increased the possibility of REEs input into the environment11. Furthermore, REEs exploitation and mining activities not only result in the degradation of the original ecosystem, severe soil erosion, and biodiversity loss, but also contribute to the spread of REEs to surrounding areas12. Rainfall accompanied by natural leaching exacerbates damage to soil structures and nutrient loss. REEs exploitation has been one of the principal factors in soil degradation in southern China.

Previous studies have already emphasized the irreversible and severe adverse effects caused by mining activities on ecological environments9. Currently, research focuses on the increase of REEs contents in the surrounding soil and water, the impact of REEs mining on crops, and the transfer of REEs between soil and plants. The influence of REEs exploitation on soil properties and microorganisms in and around mining areas have been investigated as well13,14. The concentrations of REEs in soil and water near REEs mining areas have been proven to be significantly higher than the background values of soil REEs in Jiangxi Province and China, and REEs in main crops are far higher than the critical values harmful to human health11. Researchers reported that the potential anthropogenic sources in the soil around REEs mines are composed of mining-related leachate, traffic exhaust, and industrial dust15. Moreover, mining activities decreased the content of total rare earth elements (TREEs) and altered the vertical distributions of LREEs/HREEs and chemical speciations in weathering profiles16. The relevant researches were summarized in Table S1.

The total heavy metal concentration is a common and useful indicator of soil pollution in ecosystems. However, environmental problems driven by high heavy metal levels are more strongly related to their chemical speciations and bioavailability rather than to their total contents17,18. Previous studies have reported that mining activities not only translocated heavy metals into soils19but also significantly increased their potential bioavailability in mine tailings20,21,22. This exacerbates the threat to deep soil and the surrounding environment. Owing to the unique forms of ionic REEs, the REEs morphological changes caused by mining activities should directly affect the pollution characteristics in and around mining areas, as well as the corresponding environmental risks. However, little is known about the REEs distribution and speciation differences between unexploited and mining areas that might be influenced by mining activities. Hence, more attention should be given to the transformation of REEs fractions driven by anthropogenic exploitation, as relevant research is critical for assessing the ecological risks of REEs exploitation and developing targeted environmental management strategies.

Therefore, we investigated the morphology and geochemical characteristics of REEs in unexploited and mining areas in southern China. The main purposes of this study are to (1) clarify the differences in REEs distribution characteristics and chemical morphological patterns between mining soil and unexploited soil; (2) investigate the relative importance of REEs speciation and the physicochemical properties of soils on its bioavailability. This study could provide new insights into the environmental risks associated with ionic REEs exploitation.

Materials and methods

Study area and sampling collection

The study areas were located in Longnan and Dingnan Counties, southern Jiangxi Province, China. The four key factors, including REEs-rich granites, intense chemical weathering driven by high temperature and rainy climate, clay-rich weathering profiles with high adsorption capacity, and gentle topography allowing profile preservation, create the perfect conditions for forming ionic REEs deposits in this region. This area has a subtropical monsoonal humid climate with high precipitation and mild weather. The annual average precipitation is 1500–1600 mm and the mean temperature is 19.0 ℃. The study area is mountainous, which is famous for its abundant rare earth resources, with hills and mountains accounting for nearly 80% of the region. The dominant vegetation consists of coniferous pines and shrubs, and the soil is classified as red soil. Over the past several decades, in-situ leaching has been the primary method for REEs extraction. However, all mining activities have been restricted since 2016. In abandoned mining areas, most of the vegetations and topsoil have been removed to lay plastic pipes for transporting leaching solution. Additionally, many abandoned leaching holes remain visible across the mining areas, serving as clear indicators of prior mining operations. In September 2020, three sampling areas marked as LD, LE, and LN were arranged in abandoned mining areas without any soil remediation. The sampling sites were arranged via the systematic random distribution method and adjusted according to the actual geographical situation. Approximately 1000 g topsoil (0–15 cm) was collected at each sampling site, and each sample was a mixture of four subsamples randomly collected in a circle with a radius of 4 m, Forty mining soil samples (LD1-LD12, LE1-LE15, LN1-LN13) were randomly collected individually and placed in an independent sterile polyvinyl chloride bag. Another twelve samples, marked as LA1-LA12, were collected randomly using the same method in unexploited REEs mine areas, which were in another ore district and had not undergone any forms of extraction. A total of 52 topsoil samples were gathered by a stainless steel core soil sampler, and their geographic coordinates were recorded via a global position system, as illustrated in Supplementary Table S2 and Fig. S1.

Soil property analysis and REEs characteristic parameters analysis

Visible plant residues and roots were removed, and then each sample was divided into two parts. One part was preserved at 4 °C until the assessment of soil moisture content, organic matter, pH, ammonium, nitrite, and nitrate.Moisture content was represented by the difference of weight before and after drying at 105 °C for 24 h. Organic matter was analyzed by measuring the weight difference after combustion at 550 °C for 2 h. Soil pH and electrical conductivity (EC) were measured after shaking the soil water (1:2.5 w/v) suspension for 1 h, using a pH meter (Delta 320, Mettler Toledo, Switzerland) and an electrical conductivity meter (sensIONTM5, HACH, USA), respectively. Inorganic N (nitrate, nitrite and ammonium) was extracted with 2 M KCl (1:5 of soil: solution by mass) by shaking (1 h, 200 rpm) and filtering. The filtrate was sampled to determine by spectrophotometric method. The other portion was air-dried at ambient temperature and sieved through a 2 mm polyethylene mesh for REEs analyses. Some of the air-dried soil was ground into fine powder using an agate mortar and then passed through a 0.075 mm nylon sieve for total carbon (TC), total nitrogen (TN), and total sulhur (TS) analysis, using an Elemental analyzer (Vario MAX CNS, Elementar Analysesysteme, Germany).

TREEs are composed of elements La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, and Y. LREEs are the sum of La, Ce, Pr, Nd, Sm, and Eu. HREEs are the sum of the Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, and Y. BCR (Bureau Communautaire de Référence) sequential extraction was conducted to investigate the different chemical speciations of REEs in each soil sample: exchangeable fraction (F1), reducible fraction (F2) that bound Fe/Mn oxides, oxidizable fraction (F3) that bound to organic matter, and residual fraction (F4)16. The BCR procedure was described shown in the Supplementary Information (Fig. S2). The concentration of REEs in the extracted solutions of each step was ultimately measured by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 8800 Triple Quadrupole, Agilent Technologies, US). The limit of detection (LOD) was 0.01 µg/L, and the limit of quantitation (LOQ) was 0.03 µg/L. The major elements Fe and Al were also analyzed. A calibration curve was prepared from a standard solution for each element, and the linear standard curves were better than 0.999 for all elements. To determine the precisions of the BCR method, all the samples were extracted in triplicate, and the relative standard deviation (RSD) in the fraction of F1 to F3 was below 6% on average. The contents of F1, F2, F3, and F4 were calculated based on the sum of 15 exchangeable, reducible, oxidizable, and residual REEs fractions, respectively. Fractionation characteristics of LREEs and HREEs were evaluated using LaN/SmN, GdN/YbN, and LaN/YbN ratios to illustrate their geochemical patterns23. An anomaly of REEs indicated their enrichment or depletion relative to neighboring REEs, and Ce(Ce/Ce*) and Eu(Eu/Eu*) could be calculated using the following Eqs24,25..

\(\delta Ce=\frac{{C{\text{e}}}}{{C{\text{e*}}}}{\text{=}}\frac{{C{{\text{e}}_N}}}{{{{(L{a_N} \times P{r_N})}^{0.5}}}}\)\(\delta Eu=\frac{{Eu}}{{Eu{\text{*}}}}{\text{=}}\frac{{E{u_N}}}{{{{(S{m_N} \times G{d_N})}^{0.5}}}}\)\(\mathop {\left( {La/Sm} \right)}\nolimits_{N} =L{a_N}/S{m_N}\)

\(\mathop {\left( {La/Yb} \right)}\nolimits_{N} =L{a_N}/Y{b_N}\)\(\mathop {\left( {Gd/Yb} \right)}\nolimits_{N} =G{d_N}/Y{b_N}\)

Statistical analysis

Statistical analyses were conducted by SPSS software (version 19.0) to explore the relationships between REEs and soil properties. An independent sample t-test (Student t-test or Welch’s t-test) was applied to test the differences in the actual content and proportion of the four REEs fractions between unexploited soils and mining soils. To validate the assumption of homogeneity of variances for the t-test, we performed Levene’s test. ArcMap 10.0 was used to generate a sampling distribution map. The mental test reflected the relative effects of environmental variables on REEs speciation in different soils, and displayed correlations using R software(version 4.1.2). Line graphs, Pearson correlations, and biplots were applied, and the figures above were drawn via Origin 2018.

Results

Characteristics of REEs chemical speciation patterns

Chemical speciation analysis of soils from 52 sampling points within mining areas and unexploited areas was carried out via the BCR sequential extraction method. The extracted morphological forms included exchangeable, reducible, oxidizable, and residual fractions (Fig. 1 and Fig. S3).

Fig. 1
figure 1

The proportion of REEs fractions in all the soil samples.

Full size image

TREEs in unexploited soil (LA) range from 590 to 780 mg/kg. The mean percentage distributions of the REEs four fractions in unexploited soils showed similar trends as follows: exchangeable (F1) < reducible (F2) < oxidizable (F3) < residual (F4). F1 presented the lowest content from 1.11 to 85.95 mg/kg, accounting for 0.15–13.02% of TREEs. F4 exhibited the highest content and accounted for 50.85–91.41% of TREEs.

Remarkable discrepancies in REEs contents existed among the different sampling points in the mining areas. The TREEs content of mining soils (LD, LE, and LN) varied over a broad range from 100 to 790 mg/kg, with a median of 373.25 mg/kg. The average percentage of each fraction showed a consistent trend following the gradient F3 < F1 < F4 < F2. The proportion of exchangeable fraction (F1) represented 3.79%−48.71% of TREEs, which was significantly greater than that in unexploited soils (Fig. 2). While the proportion of residual fraction (F4) notably decreased and ranged from 32.82 to 209.76 mg/kg (4.62%−60.08%), and both the actual content and proportion were extremly and significantly lower than those of unexploited soil. The contents of reducible fraction (F2) in mining soils were significantly greater than that of unexploited soils, accounting for 21.15–69.33% of TREEs.

Fig. 2
figure 2

The statistical comparisons of actual content (a) and proportion (b) of REEs fractions in unexploited soil and mining soil (* p < 0.05, ** p < 0.01, *** p < 0.001).

Full size image

Distribution pattern of REEs fractionation

In mining soils, LREEs varied from 41.65 to 191.25 mg/kg, and HREEs ranged from 43.30 to 618.90 mg/kg. Unexploited soils contained higher contents of LREEs than HREEs, whereas the mining soil samples exhibited more HREEs than LREEs. Furthermore, both mining and unexploited soils complied with the oddo-Harkins rule, which illustrated that elements with even atomic numbers (Ce, Nd, Sm, Gd, Er, and Yb) were more abundant than those with adjacent odd atomic numbers (La, Pr, Eu, Tb, Ho, Tm and Lu) (Fig. 3). As the highest content of HREEs, Y accounted for 15.28% and 51.71% of TREEs in the mining soils and unexploited soils, respectively. Due to their high content, HREEs accounted for over 70% of TREEs in mining soil samples. Ce was the most abundant LREE, ranging from 15 mg/kg to 260 mg/kg. In unexploited soil, the proportion of Ce within TREEs was significantly higher than that in mining soil, which led to a greater proportion of LREEs in unexploited soil (75.39%) than in mining soil (25.81%). These two elements represented more than 50% of TREEs in all the samples.

Fig. 3
figure 3

The contents (a) and percentages (b) of REEs species in the four sampling sites (LA, LD, LN, LE).

Full size image

The normalized REEs distributions in mining soils were consistent, with the curves of LN, LD, and LE downwards from right to left. While the normalized curve of unexploited soil (LA) tilted from left to right, and exhibited steep in La-Eu and flat in Dy-Lu (Fig. 4). Bivariate plots of REEs fractionations revealed different fractionation characteristics of LREEs and HREEs in unexploited soils and mining soils (Fig. 5). The range of LREEs/HREEs variation within unexploited soil (LA) spanned from 2.25 to 4.16, with an arithmetic mean of 3.20, which were much higher than those of mining soils. These samples also exhibited a LaN/YbN spectrum spanning from 6.91 to 14.11, demonstrating a conspicuous divergence between light and heavy REEs. Conversely, the LREEs/HREEs values in most mining soil samples were less than one, falling within the interval of 0.15 to 1.63. The values of LaN/YbN notably deviated downwards from those of unexploited soil, indicating a comparatively less pronounced partitioning between LREEs and HREEs in the mining area. The LREEs fractionation assessed by LaN/SmN (mean of 0.64 in mining soil and 4.37 in unexploited soil) was significantly lower than the HREEs fractionation calculated with GdN/YbN (mean of 0.77 in mining soil and 2.03 in unexploited soil). Both the LA and LD samples exhibited positive anomalies, with δCe values exceeding unity (δCe > 1). However, in the LN and LE samples, a portion of data points manifested positive Ce anomalies, and the others displayed negative Ce anomalies (δCe < 1). In contrast, all the samples exhibited negative Eu anomalies, as characterized by δEu values ranging from 0.05 to 0.37, with an average of 0.14.

Fig. 4
figure 4

Chondrite-normalized REEs distribution patterns in the unexploited soil (LA) and mining soil (LD, LN, and LE).

Full size image
Fig. 5
figure 5

The variation of TREEs and LREE/HREE (a), Eu/Eu* and Ce/Ce* (b), LaN/YbN and LaN/SmN (c), GdN/YbN and LaN/SmN (d) in the unexploited and mining soil samples.

Full size image

Both LREEs/HREEs and Eu/Eu* displayed extremely significant and positive correlations with LaN/YbN, LaN/SmN, and GdN/YbN (p < 0.001), respectively. A significantly positive correlation was also detected between LREEs/HREEs and Eu/Eu* (p < 0.001). The Ce/Ce* ratio was significantly and negatively correlated with GdN/YbN, indicating that an increase in GdN/YbN promoted the consumption of Ce/Ce*. pH was negatively correlated with LREEs/HREEs, Eu/Eu*, LaN/YbN, and LaN/SmN. However, organic matter was positively correlated with these variables (Fig. 6).

Fig. 6
figure 6

Correlogram among the fractionated REEs and soil properties in all the soil samples.

Full size image

Effects of soil properties on REEs characteristics

Both the mining soil and unexploited soil samples were strongly acidic, with all the pH values below 6.0. There was more organic matter and moisture content in unexploited soil than those in mining soil. However, there were no significant differences in inorganic nitrogen, TC, TN, and TS between unexploited and mining soil samples. The physicochemical properties of the 52 soil samples were shown in Table S3.

The relationships between REEs and the soil physicochemical properties were depicted in Fig. S4. Extremely significant negative correlations (p < 0.01) were observed between REEs (TREE, LREEs, and HREEs) and soil organic matter, as well as TN, in unexploited soil. In mining soil, the correlations between the REEs and soil pH were all significantly positive (p < 0.01), and no evident correlation between organic matter and REEs was detected in the mining area (p > 0.05).

The morphological forms of REEs were influenced by different soil properties in unexploited and mining areas, as reflected by the Mantel test (Fig. 7). In unexploited soil, organic matter, and TN appeared to have significant influences on the residual fraction (F4) (p < 0.01). However, in mining soils, pH was found to significantly affect all four fractions of REEs (p < 0.05), and organic matter displayed a significant correlation with exchangeable (F1) and residual (F4) fractions (p < 0.05). The moisture content was also significantly correlated with exchangeable (F1) and residual (F4) fractions.

Fig. 7
figure 7

Pairwise correlation matrix was shown with Mantel and partial Mantel test of REEs fractions with soil properties in unexploited soil (a) and mining soil (b). The order of environmental attributes was sorted based on the Pearson’s correlation coefficients with REEs fraction by partial Mantel tests, the size of the square indicated the value of coefficient between the environment factors. OM, organic matter; F1, F2, F3 and F4 refer to REEs exchangeable, reducible, oxidizable, and residual fraction, respectively.

Full size image

Al and Fe chemical fractionation

The morphological distributions of Fe and Al, obtained through BCR sequential extraction, were depicted in Fig. S5 and S6. Total Al contents in mining areas and unexploited soils were 49.36–118.18 g/kg, and 44.59–109.01 g/kg, respectively. Total Fe contents in mining areas and unexploited soils were 13.22–56.87 g/kg, and 4.15–19.77 g/kg, respectively. Across all the soil samples in unexploited and mining areas, the contents and proportions of Fe and Al exchangeable fraction were lowest. These two elements were both dominated by the residual fraction, accounting for 91%−99% and 93–99%, respectively, which was significantly greater than the other fractions. The relationships between Fe, Al, and the soil physicochemical properties were shown in Fig. S7.

Discussion

Effect of mining on the geochemical behavior of REEs fractionations

Currently, there are no soil contamination or risk standards established for REEs. However, the TREEs contents of unexploited and mining soils significantly exceed the background levels of the soil in Jiangxi (211 mg kg−1), China (188 mg kg−1)26, and the world (154 mg kg−1)7. This indicated that the surface soils in both unexploited and mining areas were enriched with high abundances of REEs. The enrichment of ionic rare earth deposits in South China is affected by several factors, such as the original parent rock, pH, weathering degree, geomorphology, particle size, clay mineral content, and soil organic matter27. Compared with unexploited soils, most mining soils presented lower TREEs contents. However, the TREEs of certain mining soils were remarkably higher than those of unexploited soils, which probably resulted from the drilling of injection holes in the ore body and covering it onto the surface soil.

The chemical morphological characteristics of REEs were investigated to better understand their geochemical patterns and environmental behavior. All the results of the normalized REEs fractionation patterns reflected an obvious enrichment of LREEs and significant LREEs to HREEs fractionation in unexploited soil. The LREEs/HREEs values of unexploited soil were higher than those of mining soils (Fig. 5a). According to the previous study, nearly 90% of the known REEs deposits are enriched with LREEs, with LREEs oxides accounting for more than 50% of TREEs resources6. In the ionic REEs mineralization process, the upper weathering crust is conducive to LREEs enrichment, as LREEs tend to accumulate in surface soils due to organic matter and Fe-Mn oxide adsorption28,29. At the same time, the disparity between mining and exploited soils in LaN/YbN and GdN/YbN values underscored the marked enrichment of HREEs over LREEs in mining areas after exploitation, especially the content of yttrium (Figs. 3a and 5c and d). HREEs were more likely to form bicarbonate and organic complexes, and then were more easily leached by the water. Thus, the middle and lower layers favor HREEs enrichment30,31. Before the leaching process, injection holes need to be excavated in the ore body. Therefore, the overturned middle and lower layers cover the surface soil, resulting in an increase of HREEs content in the mining topsoil. Moreover, yttrium oxides exhibit greater stability than oxides of other HREEs, and are thus less influenced by dissolution, precipitation, or other chemical reactions. The leaching process also significantly decreased the contents of LREEs (La, Ce, Nd, etc.) (Fig. 3). There findings demonstrated that mining operations changed the distributions of LREEs and HREEs in the surface soil of mining areas.

Unlike other lanthanides, Ce and Eu exhibit distinct valence states under varying redox conditions, subsequently altering their properties and abundance during soil formation32,33. Eu was found to be more mobile than the other REEs in the soils, whereas Ce was less mobile34. Consequently, Ce and Eu are fractionated from other REEs, leading to anomalies. Positive Ce and negative Eu anomalies are universally observed in the surface soil34. The δCe values in LA are all positive anomalies, some soil samples in the mining areas exhibit positive anomalies, whereas others display negative anomalies. These findings are consistent with most previous studies on the distribution of REEs in tailings and surrounding soils of ionic REEs deposits35,36. During supergene processes, distinct from other REEs, dissociated cerium rapidly oxidizes to CeO2. CeO2 does not migrate with surface water, but rather deposits in the humus layer and the upper part of the weathering crust, and then accumulates in the ore deposit and soil37. In addition, the soil type of REEs mining area is red soil. During red soil formation, hot and rainy conditions led to desilicification and allitization, which resulted in high contents of iron and aluminium oxide in the soil38. High iron contents in soil can also lead to the enrichment of Ce31. The negative Ce anomalies in some mining soil samples indicated a decrease in cerium content, which could be attributed to intense leaching activity and potentially result in severe REEs contamination in surrounding farmlands34,39. All the soil samples exhibited negative δEu anomalies. Eu can be easily reduced from Eu3+ to Eu2+, and can migrate downwards, generating a negative Eu anomaly30.

Additionally, positive correlations between Eu/Eu* and LREEs/HREEs, as well as with LaN/YbN, GdN/YbN, and LaN/SmN suggested that increases in these ratios could benefit the formation of Eu/Eu*. The negative correlation between Ce/Ce* and GdN/YbN indicated that an increase in GdN/YbN favoured the consumption of Ce/Ce* (Fig. 6). Consequently, fractionated REEs can induce further fractionation of other REEs40.

Fractionation of REEs and their relationships with soil properties

REEs chemical speciation and bioavailability play pivotal roles in risk and hazard assessment, ecological restoration, and the management of REE-contaminated areas10. Chemical speciations typically determine their solubility and biological availability, which are influenced by various factors such as pH, salinity, and the presence of anions41,42. Before in-situ leaching, the TREEs contents of unexploited soil were higher than those of mining areas, but the residual fraction accounted for the largest proportion, and the proportion of exchangeable fraction was the smallest (Figs. 1 and 2), which implied little impact on surroundings. The residual REEs are usually present within the soil lattice of silicates, primary minerals, and secondary minerals, which are difficult to release and extract under natural conditions, and are also difficult for biota to obtain43. However, under strongly acidic conditions, they may migrate. During the rare earth mining process, high-concentration leaching solutions remain in the soil for an extended period, reducing the stability of residual REEs lattice structures. These structures become looser under the influence of weathering, leading to the release of REEs. These REEs are then fixed by carbonates, iron and manganese oxides, and organic matter in the form of reducible and exchangeable forms44. The high contents of labile REEs in mining areas were related not only to incomplete recovery of the leaching solution, but also to the presence of high concentrations of ammonium sulfate (2–4%)16. Soils with acid sulfate in mining areas are a source of mobile and active REEs, leading to the release of REEs45. Therefore, after in-situ leaching, the residual fraction remarkably decreased, and the contents and proportions of the exchangeable and reducible fractions in mining soils significantly increased. The significant correlations between exchangeable (F1), reducible (F2), and oxidizable (F3) fractions in mining soils reflected their geochemical similarity, and these fractions are also generally considered to possess potential mobility and biological availability under various environmental conditions46. Exchangeable fraction REEs (F1) is characterized by weak sorption and bioavailability, so it is most likely to be released into the environment through rainfall and surface runoff, and cause environmental hazards and great impacts on the surroundings. Liang reported that in unexploited areas, the exchangeable fraction of REEs accounted for 14% of TREEs, but in mining areas, this proportion increased to 44–52%40. These results demonstrated that human mining activities altered the soil texture and oxidation conditions, and disrupted the capacity of the soil to fix and release metals, thereby leading to changes in soil REEs forms and increasing REEs instability.

The weak correlation between soil properties and REEs chemical fractions of unexploited soil suggested that soil properties have minimal influence on the distribution and adsorption of REEs. Nevertheless, after in-situ leaching, more complex correlations between them were found in mining soils. These results implied that exploitation has disrupted the stability of REEs chemical speciations in mining areas, and the soil properties significantly affect distributions of different REEs fractions. The distribution and behaviour of REEs are influenced by soil properties, such as parent material, weathering processes, soil formation, climate, pH, redox conditions, and organic matter47,48,49. pH plays a crucial role in controlling the mobility of REEs in environments, and exhibited a significant negative correlation with the migration rate of REEs. Because a relatively high pH significantly enhances the absorption of REEs, thus reducing their migration rate39. In the unexploited soils of this study, no significant correlation was detected between pH and REEs, or between pH and different chemical fractions of REEs. Similar results have been reported in the soils of coal mines in Brazil and other regions39,50. In the mining area, pH was significantly and positively correlated not only with TREE, but also with four different chemical fractions of REEs. This discrepancy might arise from the small pH variation range (4.1–5.3) in mining soils, which were strongly acidic and exhibited limited variation. This pH range is similar to that of red soils in southern China, where over-acidified soil conditions may mask the influence of soil pH on the distribution and behavior of REEs34.

The combinations of REEs and soil organic matter are influenced by the type, composition, and content of organic matter, as well as the soil pH and redox conditions51. Organic matter exhibited a significant negative correlation with REEs in unexploited soil, possibly due to the formation of complexes/chelates52. The low organic matter content in mining soil may lead to weak adsorption of REEs, so there is no significant correlation was detected. Moisture is a driving force for REEs displacement, as increased humidity enhances microbial activity53disrupting rare earth phosphate compounds and promoting the release of REEs54. The significant correlations between moisture content and F1/F4 would be related to the REEs chemical morphological changes caused by exploitation. In the ammonium sulfate leaching process, potential leakage of impermeable layers, inadequate collection systems, and high nitrogen contents not only results in the retention of a significant amount of nitrogen, but also traps sulfate ions in soil, concealing the relationship between TN and REEs55. Furthermore, there were extremely high outliers in TN and TC contents in some of the soil samples (e.g., TN values of LE1 and LN13, TC values of LD7 and LD8). This may be due to the presence of bird feces or calcite (CaCO3), dolomite [CaMg(CO3)2], or other carbonate minerals in the soil samples.

Effects of mining on Fe and Al morphology

Jiangxi is known as the “Red Soil Region” due to its iron-rich soil, which contains iron-aluminum oxides or carbonate substances. The soil in this region is slightly acidic, primarily because of the transformation of Al into active forms. Aluminum and iron ions are major impurities in leaching solutions of ion-absorbed rare earth ores56. Therefore, the soil samples contained high contents of Fe and Al elements. The predominant chemical speciation of Al and Fe elements in all the soil samples from both the unexploited and mining areas was the residual fraction, which was more stable, less reactive, and significantly more abundant content than the other three fractions. The predominant residual fraction and lower content of labile fractions might result in no significant correlation between the fractions of Fe/Al elements and the fractions of REEs.

Effect of REEs on the environment of mining areas

Soil formation processes and human activities significantly influence REEs concentrations and speciation. In undisturbed environments, the residual REEs dominate, followed by the oxidizable and reducible fractions, with the lowest content of exchangeable REEs. These distributions primarily result from pedogenesis processes and parent rock weathering16,57which control REEs release and fixation. Although the TREEs content may be high in unexploited soil, their limited mobility minimizes environmental and human health impacts. Mining activities alter soil physicochemical properties and redox conditions, enhancing REEs mobility and bioavailability. Although TREEs concentrations often decrease after mining, the content and proportion of exchangeable and reducible REEs increase. Elevated REEs exchangeable fraction may inhibit microbial activity and soil enzyme functions58disrupting critical biogeochemical cycles. During the process of plant restoration, REEs in bioavailable forms can uptake by plants, potentially disrupting nutrient absorption and causing oxidative stress59resulting in slow growth. Moreover, the increased bioavailability of REEs poses ecological risks through leaching or runoff, and increases human exposure through contaminated crops, dust inhalation, or water sources60,61which could cause lung and hepatocyte toxicity, and accumulation toxicity7.

Traditional monitoring focusing solely on TREEs content significantly underestimates risk in mining areas. So, routine monitoring programs in and around mining areas should mandatorily include speciation analysis (e.g., BCR sequential extraction) to quantify the bioavailable (exchangeable) and potentially mobilizable (reducible) fractions. This provides a true picture of environmental hazard. In the process of vegetation restoration in mining areas, targeted selection of plants that can adsorption/fix REEs will reduce their bioavailability and mobility. Our findings provide the critical scientific basis for designing effective remediation strategies, improving monitoring methods, and guiding mine waste management. We also emphasize the necessity of post-mining remediation (e.g., plant stabilization, soil amendment) and long-term monitoring to mitigate ecosystem risks and protect human health.

Conclusion

This study systematically analyzed the contents, distributions, and chemical speciations of REEs in unexploited soils and mining soils in southern Jiangxi Province. The results revealed presented positive Ce and negative Eu anomalies in all the soil samples, with lower TREEs contents in mining soils compared to unexploited soils. Mining activities not only reduced the content of TREEs, but also altered the LREEs/HREEs distributions and chemical speciations in mining areas. The residual fraction dominated in unexploited soils, and the proportions of the exchangeable fraction were the lowest. Mining activities significntly decreased the contents and proportions of the residual fractions, while increased the exchangeable and reducible REEs. This alteration enhanced REEs instability, mobility, and bioavailability, probably becoming potential environmental risks. These findings demonstrate that mining activities have the potential to modify the distribution and geochemical behavior of REEs, and highlight the need for stricter environmental management in mining areas to mitigate REEs contamination. However, the mechanisms driving REE fractionation and speciation changes during mining process require further investigation. Remediation strategies and regulatory measures should be explored to protect ecosystems and human health.