Behavior of U, Th and Ra isotopes in soils during a land cover change
Résumé
Over the last decades, the U-and Th-series nuclides were successfully used to determine weathering rates in various environments. The objective of this study is to assess the potential impact of the vegetation change on the U- and Th-series signal recorded in forested soils. This study was carried out from the experimental forest site of Breuil-Chenue (Morvan, France) developed by the INRA-BEF team. The native forest of the site was partially clear-felled in 1976 and replaced by monospecific plantations stands (Oak and Douglas fir). U- and Th-series disequilibria were measured in 2011 in the podzolic soils developed under the native forest, and in the two replanted stands. Separation of primaryminerals (biotite, muscovite and perthitic feldspar) and selective extractions of the Fe and Al oxides were performed to investigate the distribution of U and Th among these soil fractions.
The selective extractions suggest that a significant part of U and Th is primarily held by Fe-bearing silicates. Our results suggest that the tree substitution seems to produce a large dissolution of these minerals under the Oaks, resulting to a release of U and Th. However, below 25 cm no impact of this release was observed on U-series disequilibria. A scenario allowing to reconcile the significant mobilization of U and the constancy of U-series disequilibria is proposed. Above 25 cm, additional pedogenic redistribution of U and Th isotopes occurs in all the profiles, inducing some discrepancies between U-series disequilibria. A clear correlation between the (Th-230/U-234) ratios and the proportions of amorphous and interlayer Al hydroxides has been highlighted. This correlation suggests a mobility of U and Th isotopes strongly associated to Al dynamics in these soils rather than Fe, despite the primary location of U in the Fe-bearing silicates and the overwhelming reported control of U-VI by Fe-oxides in oxidized environments. These pedogenic processes make the shallowest horizons of podzolic soils unsuitable for U-series dating. In contrast, a soil production rate can be deduced from the deepest soil layers which do not show such effects on the U-series nuclides. The reproducible U-series disequilibria measured in four whole-profile replicates emphasize the robustness and the significance of the "long-lived" U-series disequilibria in deep soil layers relative to long-term weathering rates, independent of transient perturbations such as land cover changes. Finally, because Ra can strongly accumulates in plants, the (Ra-226/Th-230) ratios in the different soils were affected by the flux of Ra-226 released by litter degradation. The use of this ratio as a long-term chronometer should therefore be performed with caution in such contexts. No direct impact of the vegetation type on the (Ra-228/Th-232) ratios was identified, due to the short Ra-228 half-life.