COS 193-6 - Willow aboveground and belowground traits can predict phytoremediation services

Background/Question/Methods

The increasing number of contaminated sites worldwide calls for sustainable remediation techniques, such as phytoremediation, in which plants are used to decontaminate soils. However, current phytoremediation approaches often lack predictability, partly because of our poor knowledge of the plant traits associated with decontamination efficiency. Better anchoring phytoremediation in plant ecophysiology may improve plant selection and phytoremediation success. To use plant traits in phytoremediation, it is essential to demonstrate covariations with various key phytoremediation services. Therefore, traits-services correlations were investigated in willows, a diversified and polyvalent genus (Salix sp.) commonly used in phytoremediation.

This study explored how the community composition, diversity and coppicing treatment of willow plantations, along with their concurrent effects on community-level plant functional traits, influenced phytoremediation services. Based on the results of a four-year phytoremediation field trial using willows to remediate a brownfield contaminated by trace elements, the efficiency of ten plant traits for predicting six phytoremediation services was assessed. Linear mixed models were used to evaluate the effect of diversity (one or four cultivars) or coppicing on the traits and services, while a redundancy analysis tested the strength of the traits-services correlations.

Results/Conclusions

Neither willow diversity nor coppicing treatments had a consistent impact on phytoremediation services directly. These services were instead explained by willow traits related to resource economics and allocation. Trace element accumulation in the biomass correlated well to a conservative resource strategy along the plant “fast–slow” continuum. Greater belowground investments also promoted plant bioconcentration and soil decontamination. These traits–services correlations were consistent for several trace elements investigated, suggesting high generalizability among inorganic contaminants.



Overall, the study provides evidence that traits can be used as predictors in phytoremediation for a broad variety of contaminants, even with a short taxonomic (and thus functional) plant gradient. This suggests that a trait-based approach represent a promising avenue for developing predictive plant selection strategies in phytoremediation trials, through a better rooting of applied sciences in fundamental plant ecophysiology. Our results also highlight the potential of this approach for many other plant-based engineering technologies.