AbstractQuestions(1) Do species richness, flower functional diversity and redundancy of tree and non‐tree species increase with restoration age; (2) are the flower traits of the reference forest similar to those found in restored sites; and (3) does species originality, in terms of unique combinations of different set of flower traits, differ among sites?LocationForests restored by planting a high diversity assemblage of tree species, within the Atlantic Rain Forest, Brazil.MethodsWe sampled all reproductive individuals (353 species) and classified them in terms of floral type, colour, size, generalization level and openness. As most trees were planted and non‐tree species (sub‐shrub, shrub, herbs, epiphytes, climbers and hemi‐parasitic plants) were mostly naturally established, we analysed them in separate analyses. We compared species richness, functional diversity, redundancy and originality among sites. Moreover, we identified optimal subsets of flower traits to describe vegetation succession trends.Results(1) Flower functional diversity of tree species could be achieved after two decades of restoration, but remained far from reference values for non‐tree species. The same gap regarding reference values occurred with species richness, which increased over restoration periods, although were far from the reference values. Redundancy and functional diversity are not related to restoration age. (2) Red/burgundy colour, gullet and bell‐funnel shape and large size were identified as indicators of the reference forest flowers. Different indicator traits were found for each restoration sites. (3) Sites were similar regarding originality for tree and for non‐tree species.ConclusionA highly diverse species pool seems to promise to increase richness and functional diversity of tree species, but not of non‐tree species. Trait analyses have important implications for restoration projects and can be used instead of species taxonomic identity, especially for highly diverse tropical forests. Such functional diversity may be advanced through a priori definition of which species from the regional pool can be used in plantings for forest restoration, with special attention to functional traits of non‐tree species. Otherwise, restoration sites in highly fragmented landscapes will not reach functional diversity of reference sites, even after a five‐decade period.
Read full abstract