Biodiversity effects in tree seedlings are modified by light levels across functional groups

Background/Question/Methods

Tree biodiversity has the potential to maintain the functioning of forest ecosystems, not just over space, but over long-timescales by maintaining composition through recruitment. However, for continued buffering in the face of global environmental change, the sensitivity of biodiversity-ecosystem functioning relationships to heterogeneous environments needs to be understood. Biodiversity effects across environments can be driven by different mechanisms based on the nature of interactions between species and the environment. Although biodiversity in seedling communities can potentially increase their growth and biomass accumulation, abiotic drivers like light can alter this effect through divergent effects on constituent species and functional groups, resulting in complementarity effects, selection effects, spatial insurance or a combination. We tested the interactive effects of 3 levels of species richness and 3 levels of canopy cover on the growth of seedling communities consisting broad-leaved evergreen and deciduous tropical forest species using a fully crossed manipulated experiment in the Andaman Islands, India and compared these with field observations from a long-term forest plot in the same landscape.

Results/Conclusions

We show that in the first, most crucial year of seedling establishment, species richness and light increase community biomass independently. A generalized linear model showed positive effects of the light and species richness, but no interaction effects on mean basal area (deviance=12.96, df=47, p=0.999, R2=0.41) or mean height (deviance=8.24, df=47, p=1, R2=0.48) of experimental plots. These results align with observational data from similar species in the landscape. Accounting for variation across species (as random effects in a generalized linear model), individual species on average accumulated more biomass in higher light and high diversity conditions (fixed effect R2 for basal diameter=0.16, height=0.29). We also show that overyielding in species rich seedling communities does not happen through increased selection effects (selection effects decreased with species richness treatments; ANOVA F=23.3, p< 0.001). Moreover, overyielding did not correspond to spatial insurance effects as seen in experimental grassland communities, but instead fit expectations of complementarity with non-random overyielding. Together, our results show that the potential for biodiversity to increase ecosystem functioning in seedling communities is modulated by light. Our results further understanding on the interaction of biodiversity with abiotic drivers and lays groundwork for predicting future ecosystem functioning in forested landscapes under a changing environment.