Evaluation of Soil REE Perturbations on Plant Accumulation and Distribution Patterns by Multivariate data processing

Abstract

The demand for Rare Earth Elements (REE) is constantly increasing due to their use in advanced technologies, making them emerging contaminants. Although non-essential to plants, REE can accumulate in vegetal tissues and transfer through trophic levels. This raises questions about how plants respond to changes in REE composition in the soil. The present work explores the influence of soil perturbation by REE on their distribution and accumulation in Vitis Vinifera L. plant using multivariate data processing. For this purpose, a full factorial (3²) experimental plan was designed, varying two factors at three levels: treatment (control, low, high) and position (soil, roots, leaves). In more detail, for treatment two stress levels were applied by adding equimolar REE solutions to the soil: in the “low” condition, REE concentration [REE] was ~3 times higher than in the “control” one; in the “high” condition [REE] was ~50 times higher. At full leaf development, roots, leaves, and soil were collected and REE determined by ICP-MS. The high differences in REE concentration between treatments and positions, along with the effects of equimolar addition, prevented the detection of evident correlation patterns between factors and variables. Therefore, a two-step pre-processing strategy was implemented: orthogonalization with respect to PC1, to minimize concentration effects while preserving the design structure, followed by sum-100 normalization, to minimize global intensity effects due to the equimolar addition. Subsequently, Principal Component Analysis (PCA) successfully differentiated REE profiles across treatments, enabling unsupervised differentiation of experimental conditions based on multi-elemental distribution patterns, rather than on absolute concentrations. To identify meaningful associations between factors and REE patterns, ANOVA–Simultaneous Component Analysis (ASCA) was applied. As expected from the experimental design, treatment had the strongest effect. Notably, treatment × position interactions were more significant than the position alone, indicating that REE distributions may change with treatment in a compartment-specific manner. In conclusion, the present study reveals that REE accumulation and distribution in plant tissues depend not only on total REE levels but also on their relative distribution in the soil, suggesting valuable implications for environmental monitoring, pollution assessment and geographical traceability.