In recent decades, quantitative trait-based models have been used to predict how plant communities will respond to abiotic and biotic perturbations. Species trait values from literature or large databases that are most often obtained from mature plants are typically used to inform such models. These values may not be appropriate for understanding community processes of restored systems for two reasons: (1) trait values from different populations of the same species may vary and (2) trait values of mature plants may not reflect seedling traits that are important for establishment and survival. To address these concerns, we measured traits (e.g. specific leaf area, growth rate, root to shoot ratio, leaf dry mater content, root length) of eight grass and forb species commonly used in arid land restoration from several populations at different ontogenetic stages. We then used Bayesian point estimates of population trait means and variability at each ontogenetic stage and compared them to one another as well as trait values from the TRY Plant Trait Database or values published in the literature.
Results/Conclusions
We found that mean trait values as well as population-level trait variability differed by population, trait measured, and ontogenetic stage for all of the species investigated. In some cases, differences in trait values at specific ontogenetic stages resulted in unique ordering of populations within a species. In several cases, trait values of young seedlings differed greatly from those reported in the TRY Plant Trait Database or the literature.
Because traits expressed during early stages of plant growth are critical to plant establishment, a better understanding of variation in seedling traits will inform seed source selection for restoration and improve the use of trait-based models for predicting re-vegetation outcomes.