Dynamics of Rare Earths and associated major and trace elements during Douglas-fir ( Pseudotsuga menziesii ) and European beech ( Fagus sylvatica L.) litter degradation.

. Given the diverse physico-chemical properties of elements, we hypothesize that their incoherent 15 distribution across the leaf tissues, combined with the distinct resistance to degradation that each tissue exhibits, leads to distinct turnover rates between elements. Moreover, litter layers of different ages produce diverse chemical signatures in solution during the wet degradation. To verify our hypothesis, Na, K, Mg, Mn, Ca, Pb, Al and Fe were analysed together with the Rare Earth Elements (REE) in the solid fractions and in the respective leachates of fresh leaves and different humus layers of two forested soils developed under Pseudotsuga menziesii 20 and Fagus sylvatica L. trees. The results from the leaching experiment were also compared to the in situ REE composition of the soil solutions to clarify the impact that the litter degradation processes may have on soil solution chemical compositions. Our results clearly show that REE, Al, Fe and Pb were preferentially retained in the solid litter material, in comparison to the other cations, and that their concentrations increased over time during the litter degradation. 25 Accordingly, different litter fractions produced different yields of elements and REE patterns in the leachates, indicating that the tree species and the age of the litter play a role in the chemical release during the degradation. In particular, the evolution of the REE patterns according to the age of the litter layers allowed us to deliver new findings on REE fractionation and mobilization during litter degradation. In particular, the La N /Yb N ratio highlights differences in litter degradation intensity between both tree species, which was not shown with major 30 cations. We finally showed the primary control effect that litter degradation can have on the REE composition of the soil solution, which presents a positive Ce anomaly associated with the dissolution and/or transportation of Ce-enriched MnO 2 particles accumulated onto the surface of the old litter due to white fungi activity. Similar MREE and HREE enrichments were also found in the leachates and the soil solution, probably due to their higher affinity to the organic acids, which represent the primary products from the organic matter degradation. (hereafter referred to as FL) were collected from 10 adult trees randomly selected per plot. The leaves were taken from different branches, accessible from the 140 ground, at various heights and radial directions. All leaves were aggregated together in one sample per plot and stored in clean polypropylene bags. The litter material was collected from five different locations within an area of 500 m 2 of each experimental plot using a 25x25cm metallic frame and avoiding contamination by soil particles. During the collection, the different fractions of litter were sorted according to their degradation degree (Fig. 1) and stored in polypropylene bags. In BeP, three litter fractions were identified: the new litter (OLn - unprocessed, unfragmented, light-brownish coloured), the old litter (OLv - slightly altered, bleached and softened, discoloured or dark-brownish coloured) and the fragmented litter (OF - partially decomposed and fragmented, grey-black coloured). For DoP only two fractions stood out: the new litter (OLn - unprocessed, unfragmented, light-brownish coloured) and the old litter (OLv - slightly altered, bleached and softened, grey-black coloured), whereas the fragmented litter layer 150 was not sufficiently developed to be representative as a humus layer and was not considered in this study. patterns of REE in Douglas-fir samples show an HREE enrichment when compared to the other elements of the series (0.05≤La N /Yb N ≤0.15 and 0.25≤Gd N /Yb N ≤0.51), indicating a preferential release of the heavy REE to the solution. In the European beech samples, the patterns are smoother between MREE and HREE 465 (0.94≤Gd N /Yb N ≤1.21), while they conserve the depletion in LREE when compared to the other groups (0.32≤La N /Gd N ≤0.44 and 0.38≤La N /Yb N ≤0.48). of REE towards the products of the decay as the main factor for their partitioning between the remaining solid fraction of litter and the resulting solution.

The major elements and metals can be sorted into three different groups according to the evolution of their concentrations in the bulk samples of the different litter layers (Fig. 2a-b). This classification stays coherent for Ca and Mn present a progressive increase of their concentration for the European beech, whereas they are less concentrated in the oldest litter layer of Douglas-fir in comparison to Do FL and Do OLn. The concentrations of the other trivalent metals (Fe, Al) evolve similarly to the REE with a progressive increase from the fresh leaves to the OL litters and a significant enrichment in the oldest litter layer for both species (Do OLV and Be OF). In contrast to the European beech, Douglas-fir fresh leaves present metals concentrations as high as in the Do OLn 355 litter sample.

Chemical composition, pH and DOC content of leachates
The leaching experiment led to similar REE concentration ranges between the two tree species, except for the beech fresh leaves, which are one order of magnitude less concentrated (Table SI-4). The leachate patterns present strong differences to the REE characteristics of the bulk leaves and litter material and in between the two tree 360 species.
The total REE concentrations of the Douglas-fir leachates are similar in Do FL and Do OLn samples and higher in the leachate of the Do OLv sample. As the HREE show very similar concentrations in the three leachates, the difference is mainly related to the concentration of LREE, as shown by the dust-normalized REE patterns (Fig. 3c). Indeed, LREE are similarly depleted in the leachates of the two younger samples, with LaN/YbN ratios of 0.10, 365 while Do OLv presents a LaN/YbN ratio equal to 0.82. Noticeable are the significant Eu positive anomalies (Eu/Eu*) of 1.23 and 1.31 observed in the leachates of Do OLn and Do OLv, respectively. In Do OLv a slight positive Ce anomaly (Ce/Ce*) of 1.16 is also observed.
The REE concentrations in the leachates of beech samples increase from the fresh leaves to the highest stages of litter degradation but are higher for the Be OLv material. The dust-normalized REE patterns of beech 370 leachates have an MREE enrichment in comparison to LREE and HREE where Gd shows the highest concentrations. The patterns of Be OLn and Be OLv leachates present very similar characteristics as indicated by their LaN/GdN ratios (0.46 and 0.43, respectively), GdN/YbN ratios (2.10 and 2.14, respectively) and the absence of any anomaly, whereas the Be OF leachate presents significant Ce and Gd positive anomalies (Ce/Ce*=1.49 and Gd/Gd*=1.52). In contrast, the leachate of the fresh beech leaves presents a Ce negative anomaly (Ce/Ce*=0.74). 375 Similar trends in the percentage of elements leached from the material of both tree species were observed.
The percentage of leaching of the studied elements can be classified according to their valence Na, K > Mg, Mn > Ca, Pb > Al, Fe, REE, with the trivalent elements being less leached. However, some differences can be highlighted between beech and Douglas-fir. samples. Trivalent metals and Pb, similarly to those of Douglas-fir, show a low release from the solid material during the experiment (in the case of Al, the release from fresh leaves is below the limit of quantification, as well 390 as for many REE as shown in Table SI-4 and SI-5).
The Y/Ho ratios of the leachates of Douglas-fir litter can explain the appearance of a small Y depletion in the oldest litter sample. The leachate of Do OLn, which represents the stage of degradation that brings about the formation of the OLv fraction, indeed shows a Y/Ho ratio equal to 31.29 indicating a preferential release of Y (when compared to the neighbour) that leads to a lower-than-atmospheric dust value in the Do OLv sample (Y/Ho 395 = 23.58 in Do OLv and Y/Ho = 25.93 in atmospheric dust).
In beech samples, we can observe a similar behaviour with values that are slightly higher. The Y/Ho ratio in the Be OLv leachate (Y/Ho = 36.59) justifies the absence of Y enrichment in the Be OF fraction, which instead shows a dust-like ratio (Y/Ho = 25.43).
The highest DOC concentrations (Table SI-4) were measured in the Douglas-fir leachates with values 400 ranging from 10.39 mg L -1 in Do OLv to 29.37 mg L -1 in Do OLn, while beech leachates showed concentrations from 5.91 mg L -1 in Be OLn to 14.68 mg L -1 in Be OLv. Note that for both species, the highest DOC concentrations were measured in leachates in the second-to-last degradation stages with a significant decrease in the oldest fractions.
The pH appears to be inversely proportional to the DOC concentrations. Indeed, for the Douglas-fir leachates, the 405 most acidic pH was found in Do OLn, which showed a pH = 4.26 (Table SI-4), while the pH of Do OLv was the highest with a value equal to 5.03. In beech leachates, the lowest pH was measured in Be OLv (pH = 4.07) and the highest in Be OLn (pH = 5.39).

Average REE in soil solutions
The average REE concentrations in soil solutions collected between 2012 and 2014 are reported in Table SI -6. 410 The REE total concentrations differed in one order of magnitude and were lower under the Douglas-fir stand at 20 and 40 cm depth (REE=0.88 g L -1 and 0.92 g L -1 in Do SS20 and Do SS40, respectively), whereas the highest concentration was observed in beech samples at 40 cm depth (REE=6.70 µg L -1 in Be SS40). nature has only recently emerged (Gwenzi et al., 2018). However, the toxicity of REE in plants is far beyond the scope of this work, we limit ourselves to mentioning that researchers observed REE displaying redox-related toxicity mechanisms (Hassan Ragab El-Ramady, 2010;Pagano et al., 2015) and we assume that plants can trap 505 REE in lignified tissues as a defence to avoid the toxicity-related events with the same mechanism adopted for other potentially toxic metals, such as Pb and Al. Therefore, we propose that lignins constrain the REE in the oldest litter fractions during the degradation of the leaves. Given the high affinity of such metals for oxygen, the absorption operated by lignins through the binding with the oxygen-bearing functional groups (such as phenolic, hydroxyl) may be the mechanism involved and would explain the accumulation of these metals in the oldest litter 510 fractions. Therefore, during the living cycle of leaves, lignins are able to sequestrate the elements that show higher affinity for the exposed functional groups. As lignins are the most resistant tree components in forests, they would prevent the release of the absorbed elements for longer during the litter degradation. The chemical elements that are more important for tree nutrition and metabolism would then be preferentially released to the soil solutions.
As shown by the evolution of the chemical composition of leaves and litter fractions along the different 515 degradation stages, our hypothesis on the distribution of different elements among the different tissues is confirmed. This finding is further corroborated by the result of the leaching experiment, which clearly confirms our hypothesis that during litter decay, the release of elements is linked to the degradation stage of the litter itself.
As conjectured, elements partitioned in the most labile tissues are more easily released during the degradation process than those bound to more refractory tissues, which are instead accumulated over time. 520 According to our findings, two main REE fractionation processes are specific to a leaf's life-span: (i) An inter-tissue fractionation occurring during the leave's "living period", through which recalcitrant tissues would preferentially absorb REE as a result of binding with lignins, developing a particular signature; (ii) A degradation-driven fractionation, which has the different affinities of REE towards the products of the decay as the main factor for their partitioning between the remaining solid fraction of litter and the resulting solution. 525

Cerium anomalies in leachates
Another interesting aspect of our results is the presence of small but significant positive Ce anomalies in the leachates of the oldest litter fractions (Ce/Ce*=1.16 and Ce/Ce*=1.49 in Do OLv and Be OF leachates, respectively) and a slight W-type tetrad effect in the Be OF leachate. The tetrad effect can be defined as a graphical effect that divides the REE patterns into 4 segments, so-called "tetrads" (T1 from La to Nd; T2 from Pm to Gd; 530 T3 from Gd to Ho; T4 from Er to Lu), resulting from the increased stability at a quarter, half, three-quarter and complete filling of the 4f orbital (McLennan, 1994). The tetrad effect is usually classified according to the shape of the patterns into the "W" type and "M" type. Unlike organic acids, manganese oxides are capable of a selective adsorption of Ce, along the other REE, with a mechanism of oxidative scavenging through which Ce is preferentially trapped onto the surface of the above-mentioned oxides (Bau, 1999;Bau andKoschinsky, 2009, Pourret andDavranche, 2013). The Ce enrichment linked to Mn oxides could be the reason for the formation of positive Ce anomalies in the waters that leached the litter material. 555 We conjecture that after a rainfall event, residual water that is deposited onto the surface of the oldest litter fraction has inherited a specific REE signature after passing through the younger litter layers above. We can assume that such a signature is similar to that of the leachates of the younger litter fractions recovered during our experiment (with the related MREE-HREE enrichment). Once the MnO2 deposited onto the litter surface interacts with this solution, it would preferentially adsorb Ce with the scavenging mechanism previously mentioned. A 560 question mark here is related to the form (complexed or free ions) of the REE when they enter into contact with the MnO2. We assume that their main form during such an interaction occurs mainly as free ions as their complexation with organic acids could inhibit the preferential adsorption of Ce onto MnO2 as observed by Davranche et al. (2005 and2008). They also highlighted a process of REE-organic acids complexes dissociating with time and with decreasing HA/MnO2 ratios. The reduced DOC concentrations (Table SI- Table SI-4. This strengthens the assumption that the REE patterns in leachates of fresh leaves and young litter fractions are shaped by the presence of organic acids, which confers the typical increasing trend from La to Lu and the absence of positive Ce/Ce* and of TE (Fig. 4) We propose that the process leading to Ce enrichment in waters that are in contact with the oldest litter fractions 580 occurs in three steps, as reported below (and more accurately in Figure 5

REE in soil solutions
The differences in the average soil solution REE patterns observed between the two experimental sites in the Weierbach catchment seem to be linked to the different REE release occurring during the degradation of the 720 litter at each plot. Indeed, from a depth close to the litter layers (soil solution at 20 cm depth) to the deepest soil layer (soil solution at 60 cm depth), the evolution of the HREE enrichment, Ce anomaly and specific MREE (Gd and/or Eu) enrichments in soil solutions (Fig. 7) could be discussed according to similarities with the litter leachates ( Fig. 3 c-d). It may be expected that if any litter degradation compounds can contribute to the soil solution REE composition, it would be more easily observed close to the surface and would disappear 725 progressively with depth, being diluted by the water-rock interaction processes and changing in redox conditions that control REE in soils (Braun et al., 1998;Laveuf and Cornu, 2009). Our results are in accordance with this expectation. For instance, particularities in the REE patterns of these litter leachates (especially for the last two stages of degradations for both species) seem to be mirrored by their respective soil solutions. Indeed, Eu and  It must be said that the same environmental conditions to which the litter is generally exposed are not found in the laboratory. Conditions under which a greater degradation efficiency would be expected were avoided due to limitations present in the laboratory (such as the limited exchange of gases with the atmosphere, limited light, greater volume of water per litter surface area, lower concentration of microorganisms). Additional in-situ studies 775 regarding the REE dynamics in the Weierbach catchment's soils are necessary to better understand and quantify the real contribution of litter degradation to the REE composition of soil solutions in a forest ecosystem. Legout, A., Stille, P. and Hissler, C.: Genesis and evolution of regoliths: Evidence from trace and major elements and Sr-Nd-Pb-U isotopes, Catena, 149, 185-198, doi:10.1016Catena, 149, 185-198, doi:10. /j.catena.2016Catena, 149, 185-198, doi:10. .09.015, 2017 Morris, L.A. Nutrient cycling, Enciclopedia of Forest Sciences, 3, 1227-1235, ISBN 0-12-145160-7. Nakada, R., Shibuya, T., Suzuki, K. and Takahashi