Résumé
Questions While positive effects of tree diversity on tree community biomass have often been reported in mature stands, the debate on whether diversity effects may be detectable at the seedling level persists, with opposing outcomes found so far. We still lack a comprehensive evaluation of the biodiversity effects (so-called ‘complementarity’ and ‘selection’ effects), as well as the phenotypic drivers at play, underlying early-community biomass. Even less is known about whether such biodiversity effects may change under water-limited conditions. Location Seeds from four tree species coexisting in a Mediterranean forest (Spain). Methods We built experimental tree seedling assemblages with three diversity levels — monocultures, two-species and four-species mixtures — and under two soil moisture conditions. We quantified the extent to which species richness, species identity, community-weighted mean (CWM) and functional dissimilarity (FD) influence complementarity and selection effects. We computed CWM and FD for seven functional traits related to water and light acquisition; and we calculated the complementarity and selection effects from above- and below-ground biomass measures at the community level. Results Our results showed that complementarity drove the greater biomass in more diverse assemblages at the seedling stage. This pattern was largely favored by a particular species, Quercus faginea , with distinct phenotypic traits (great height, lateral ramification and root biomass with high dry matter content), which induced a positive effect of CWM on community biomass. Moreover, our study showed that the water deficit limited the production of above-ground biomass without interacting with the community’s species richness. Conclusion Our study provides evidence that positive biodiversity effects on community biomass occur early, at the seedling stage, and it highlights the essential role that certain tree species play from their initial development stages by favoring spatial resource partitioning. Our work motivates future studies to apply integrated approaches in assessing both the community-wide and species-specific effects to understand the biodiversity–biomass relationship.