3.1 Trace elements
3.1.1 Trace element concentrations
Trace element concentrations were measured in various compartments (Table 1 and Table S1). Higher values were reported in soil samples, ranging from Cd (0.33 and 0.18 mg kg−1 at 0–30 and 30–50 cm depth, respectively) to Ti (2030 and 2606 mg kg−1, respectively). In plant samples, trace elements were, on average, 3-times more concentrated in roots compared to shoots and still ranged from Cd (on average 0.41 in roots and 0.08 mg kg−1 in shoots) to Ti (301 and 114 mg kg−1, respectively). Trace element concentrations in roots generally increased from Trifolium resupinatum to Triticum spelta (except for Cd, Cr, Cu, Sr, and Zn), while those measured in shoots generally increased from Vicia faba to Triticum spelta (except for Cu, Sr, and Zn). In soil water-extract samples, concentrations increased from Cd (0.08 and 0.04 µg kg−1 at 0–30 and 30–50 cm depth, respectively) to Ba (136 and 62.3 µg kg−1, respectively). The soil water-extract fraction represented from 0.01 to 1.9‰ of the total 0–30 cm depth soil concentrations (increasing in the following order: Ti, Cr, Co, Mn, Pb, V, Cd, Sr, Ba, Cu, and Zn) and from 0.02 to 0.75‰ for the 30–50 cm depth (increasing in the following order: Ti, Co, Cr, Mn, Pb, V, Cd, Ba, Sr, Cu, and Zn). Concentrations in soil solutions ranged from Cd (0.02 µg L−1) to Zn (160 µg L−1) with noticeable heterogeneity between sampling dates (average relative standard deviation of 47%). The soil solution/soil ratios showed the same order of elements as observed for element extractability. In general, soil solutions were more concentrated than snow sample, except for Cd, Cr, Mn, Pb, and Ti. Finally, earthworm samples showed lower concentrations than shoots, except for Cd, Co, and Zn.
3.1.2 Trace element transfer factors
Two transfer factors were calculated for each trace element and each sampling date, including soil to plant (i.e., soil to shoot, TFsoil-plant) and root to shoot (TFroot-shoot) transfer factors. The three species collected were considered: Trifolium resupinatum (Figure 1), Vicia faba (Figure S1A–B), and Triticum spelta (Figure S1C–D). The transfer of trace elements from soil to Trifolium resupinatum showed significant differences between elements, including Ti, V, Pb, Co, and Cr with the minimum values (TFsoil-plant < 0.05) and Zn, Sr, and Cu with the maximum values (TFsoil-plant > 0.5) for the first sampling date (Figure 1A). Statistically significant differences were observed between sampling dates with increasing TFsoil-plant values over time (>2 times for half of the elements, and up to >20 times for Ti), except for Cd and Sr where no evolution was observed. Zinc showed a singular pattern with higher transfer from soil to plant during the early season.
The root to shoot transfer showed more consistent pattern, both between trace elements and sampling dates (Figure 1B). Cadmium, however, appeared as the lowest transferred element. TFroot-shoot values were generally higher than TFsoil-plant ones, except for Cd, Cu, and Zn. The two other plant species showed similar patterns as described for Trifolium resupinatum for both factors, including the singular Zn pattern for TFsoil-plant (Figure S1).
3.1.3 Trace element relationships
The PCA performed to identify the relationships between trace elements explained 76% of data variance with the first two components (Figure 2A). The first component (54% of data variance) was influenced by Sr, Zn, Cu, and Ba (positive scores) and Ti, Pb, Mn, and Cr (negative scores). This dimension opposed soil solution and, to a lesser extent, soil water extract (0–30 cm) samples with positive scores and soil samples with negative scores. Also, plant samples have essentially negative scores (up to soil sample ones). The second component (22% of data variance) grouped V and Ba (positive scores) in opposition to Cd and, to a lesser extent, Co (negative scores). Earthworm samples mostly drove this dimension with negative scores, mainly influenced by Cd concentrations.
Based on the PCA, we considered contrasting elements (Cd, Cu, and Mn) to plot relationships between trace elements for all compartments: Cd/Mn vs Cu/Mn (Figure 2B). Two main end-members emerged: soil (with low Cd/Mn and Cu/Mn ratios) and soil solution (with higher Cd/Mn and Cu/Mn ratios, including a higher heterogeneity) samples. Soil water extract samples showed an intermediate signature with 30–50 cm samples highlighting lower ratio values than 0–30 cm samples. Despite the intermediate position of plant samples, they did not show a linearity between the two end-members: roots showed higher relative Cd/Mn ratios while shoots had higher relative Cu/Mn ratios. Also, the shoot signature for Triticum spelta was close to the soil–soil solution linear relationship, including samples with higher relative Cd/Mn ratios, unlike Trifolium resupinatum and Vicia faba, and the root signature showed increasing Cd/Mn and Cu/Mn ratios from Triticum spelta to Trifolium resupinatum. Finally, earthworm samples had a higher relative Cd/Mn ratio than all other compartments (from 40 to 500 times for soil solution and soil samples, respectively) and the snow sample was located between the two previously mentioned end-members.
3.2 Rare earth elements
3.2.1 Rare earth element concentrations and transfer factors
Total REE (ΣREE) showed decreasing concentrations in the following order (Table 1 and Table S1): soil (111 mg kg−1 at 0–30 cm depth) > root > shoot > earthworm (1.07 mg kg−1). Soil water-extract samples represented 0.04 and 0.12‰ of soil ΣREE concentrations at 0–30 and 30–50 cm soil depths, respectively: ΣREE concentrations were about 2.7-times higher at 30–50 cm compared to 0–30 cm. Heavy REE were 42 and 36% more extractable than LREE at 0–30 and 30–50 cm soil depths, respectively. Moreover, REE extractability was in the same range as Ti, Co, Cr, Mn, and Pb. Soil solutions had ΣREE concentrations of 0.40 µg L−1, corresponding to half of the snow sample ΣREE concentrations. The soil solution/soil REE concentration ratios showed similar trends as observed for soil water-extract/soil, with increasing ratios of 90% from LREE to HREE. However, it is notable that Ce was relatively less extractable (16 and 38% at 0–30 and 30–50 cm depth, respectively) and Eu relatively more extractable (30 and 18% at 0–30 and 30–50 cm depth) compared to their neighboring elements in the periodic table.
The transfer factor of REE from soil to Trifolium resupinatum showed similar pattern as observed for the other trace elements studied with increasing TFsoil–plant over time, despite no statistically significant difference (Figure 1A). The TFsoil–plant values measured for REE (from 0.04 to 0.14 according to the element and sampling date) were close to those of Ti, Cr, or Ba. However, distinct behaviors appeared with increasing TFsoil–plant values of 20% from average LREE to average HREE. The root to shoot transfer of REE in Trifolium resupinatum showed consistent pattern between sampling dates compared to the other trace elements (Figure 1B). Similar trends were observed for Vicia faba and Triticum spelta (Figure S1).
3.2.2 REE normalized patterns of soil and soil water-extract samples
UCC-normalized REE patterns were studied at both 0–30 and 30–50 cm soil depths (Figure 3A). Results showed similar patterns for the three sub-areas, indicating a low spatial heterogeneity. We thus averaged sub-areas in the following results. The comparison between the two soil depths also showed similar patterns. Normalized LREE/HREE ratios were >1, with statistically significant higher values (p < 0.05) in the topsoil (average LREE/HREE ratio of 1.24 and 1.18 at 0–30 and 30–50 cm soil depth, respectively). All REE patterns showed a negative Eu anomaly (on average, 0.80 and 0.82 at 0–30 and 30–50 cm soil depth, respectively).
To understand the REE transfer from soil to plant, we studied the extractable fraction considering soil water-extract samples normalized by the related soil horizon (0–30 and 30–50 cm soil depth, Figure 3B). The results showed distinct REE behaviors in soil water-extract samples from soil ones: relative MREE enrichment, LREE/HREE ratios < 1, negative Ce anomaly, and positive Eu anomaly (Figure 3B). Although soil water-extract REE patterns were similar between both soil depths, the negative Ce anomaly was more pronounced at 30–50 cm depth (0.63) compared to 0–30 cm depth (0.85).
3.2.3 REE normalized patterns of soil solution and snow samples
All soil solution samples showed similar soil-normalized REE patterns (Figure 4A), including negative Ce anomaly (on average, 0.61), negative Tb anomaly (0.64), and low LREE/HREE ratios (0.51). These characteristics were not fully observed in soil (Figure 3A) and soil water-extract (Figure 3B) samples. Only the soil solution sample collected on 15 Dec. 2020 was slightly different from the other soil solution samples, with 2.7-times higher ΣREE concentrations and 1.5-times less pronounced negative Ce anomaly.
The only snow sample collected on 14 Jan. 2021 showed distinct REE pattern to soil solution ones: no obvious negative Ce and Tb anomalies and no relative HREE enrichment (Figure 4A). We thus normalized soil solution samples to the snow concentrations to highlight the main differences between these two compartments (Figure 4B). Therefore, snow-normalized REE profiles still showed negative Ce (on average, 0.59) and Tb (0.72) anomalies and presented a low negative Yb anomaly (0.81). Beside theses anomalies, the REE profiles were relatively flat with the exception of last HREE (i.e., Tm to Lu) enrichments (LREE/HREE of 0.68).
3.2.4 REE normalized patterns of plant and earthworm samples
Three plant species were analyzed distinguishing shoots from roots. Soil-normalized REE patterns showed similar trends between considered both species and organs without noticeable anomaly when comparing to soil compartment (Figure 5A-C). Despite these relatively flat profiles, all REE patterns showed a relative MREE depletion, modifying the LREE/HREE ratio (on average 0.86 for shoots and 0.83 for roots). Rare earth element concentrations, however, highlighted systematic higher values in roots compared to shoots (on average, 6.0 times for Vicia faba, 2.7 times for Trifolium resupinatum, and 3.4 times for Triticum spelta). Also, REE concentrations increased over time in shoots (with the exception of Vicia faba for the first two sampling dates) and roots (with the exception of Triticum spelta). The increasing ratios in the three species between the first sampling date (16 Oct. 2020) and the last one (4 Feb. 2021) were 2.5, 2.7, and 4.1 for shoots of Vicia faba, Trifolium resupinatum, and Triticum spelta, respectively. The respective ratios for roots showed either lower (1.8 for Vicia faba and 2.4 for Trifolium resupinatum) or no evolution over time (for Triticum spelta).
Earthworms from each sub-area were analyzed and indicated different soil-normalized REE patterns than the other compartments considered: LREE/HREE ratios of 0.49 and negative Ce anomaly (0.69) as observed in soil solution samples, but no pronounced negative Tb anomaly (Figure 5D). One composite sample (representing a sub-area) showed singular REE pattern with higher relative HREE concentrations and more pronounced negative Eu anomaly (0.76).
3.2.5 Rare earth element relationships
As for trace elements, a PCA was performed on REE concentrations and the first two components accounted for 87% of data variance (Figure 6A). The first component (70% of data variance) was influenced by Dy, Er, Ho, Yb, Lu, and Tm (positive scores) and Pr, La, Ce, and Nd (negative scores). As observed on the first component of the trace element PCA (Figure 2A), soil solution and, to a lesser extent, soil water-extract (0–30 cm) samples (positive scores), were opposed to soil samples (negative scores). Also, the REE signatures of plant and soil samples were closer compared to the other trace elements. The second component (17% of data variance) grouped Sm, Eu, Gd, and, to a lesser extent, Nd (positive scores).
Based on these results and considering the tracer potential of some REE, we plotted the relationship between Ce/Eu (proxy of both Ce and Eu anomalies) and La/Lu (proxy of LHREE/HREE fractionation) for all samples collected in this study (Figure 6B). Once again, soil solution (with relatively low Ce/Eu and La/Lu ratios) and soil samples (with relatively high Ce/Eu and La/Lu ratios) were the two main end-members. Plant signatures were close to soil ones (despite lower Ce/Eu and La/Lu ratios). The heterogeneity of plant samples, however, was neither related to sampling dates nor plant species unlike the other trace elements. Yet, shoots showed higher relative Ce/Eu ratio than roots, resulting in a discrepancy from the linear relationship between the two main end-members. Soil water-extract samples showed an intermediate signature despite lower relative Ce/Eu ratio. Finally, earthworm samples presented a high heterogeneity and the snow sample had a signature between soil and soil water-extract samples.