Lessons from applying a trait based framework for predicting restoration success

Background/Question/Methods

Plants may have dissimilar effects on ecosystem processes because they possess different attributes.  Given increasing biodiversity losses, it is important to understand which plant traits are key drivers of ecosystem functions.  To address this question, we studied the response of two ecosystem functions that remove nitrogen (N) from wetland soils, the accumulation of N in plant biomass and denitrification potential (DNP), to variation in plant trait composition.  Our experiment manipulated plant composition in a riparian wetland.  We determined relative importance of different plant traits as predictors of each ecosystem function. 

Results/Conclusions

We demonstrate that Water Use Efficiency (WUE) had a strong negative effect on biomass N.  Root porosity and belowground biomass were negatively correlated with DNP.  Trait ordination indicated that WUE was largely orthogonal to traits that maximized DNP.  These results indicate that plant species with different trait values are required to maintain multiple ecosystem functions, and provide a more mechanistic, trait-based link between the recent findings that higher biodiversity is necessary for multi-functionality.

There were a number of challenges in taking a trait-based approach to our field ecology study.  First, while we selected plant traits based on ecological theory, several of the plant traits were not good predictors of each ecosystem function suggesting the ecological theory linking traits to function may be incomplete and require strengthening.  However, it was also very challenging to measure some of the plant traits such that we chose to grow plants in the greenhouse for destructive sampling.  Yet it is not well-understood how plastic many types of traits are and whether values will be significantly different for plants grown under different conditions.  Therefore, it may be that relationships between greenhouse-measured traits and field-measured ecosystem functions were weak due to the fact that measurements were not all completed in the field. 

In order to move the field of trait-based ecology forward, research to build up plant trait databases would be useful so that traits would not necessarily have to be measured individually for every trait-based research question.  Additionally, research is needed to understand how similar trait values are when measured on the same species but under different sets of environmental conditions.