The positive effect of diversity on ecosystem function in grassland communities is strongly supported by studies that manipulate community components. However, we need to resolve how species within a community interact to create diversity effects. The Diversity-Interactions modeling approach uses linear methods to describe the relationship between species relative abundances and ecosystem function responses. Species identity effects can be separated from species interaction effects (diversity effects), and biological hypotheses can be tested by comparing the fit of models varying in the nature of species interactions. In an effort better understand the contribution of species interactions to diversity effects, Diversity-Interactions modeling was used to analyze data collected from the first three years of the SPaCE biodiversity experiment at the University of North Dakota. Plots (1 x 1 m) were planted from a pool of 16 native grassland species and varied in species richness (monocultures, 2, 4, and 8), Simpson’s evenness (low, medium, and high), and species pattern (planted randomly or aggregated in groups of 4 conspecifics). Plots were weeded monthly, and at the end of each growing season aboveground biomass was clipped, dried, and weighed.
Results/Conclusions
In the first year of the experiment, the model with additive species-specific contributions to interactions best explained biomass production. This suggests that the diversity effect arises from each species having a fixed contribution to pairwise interactions no matter the identity of the other species in the interaction. Biomass production was driven by the forb Helianthus maximiliani, which had the greatest identity effect and only significant positive estimated contribution to pairwise interactions. In year two, the average pairwise interaction model best explained biomass production. In this model, all pairwise interactions have the same strength no matter the identity of the two species, and the diversity effect is greatest at maximum richness and evenness. In year three, again the additive species-specific interaction model best explained biomass production. Inclusion of a random term, representing individual pairwise interactions and based on species neighborhood composition, improved model fit. The warm-season grass Panicum virgatum and the legume Astragalus canadensis had significant positive estimated contributions to pairwise interactions, which suggests increased nitrogen availability and retention increased biomass. These results clarify species-specific influences on diversity effects and reinforce findings that drivers of diversity effects and species-specific contributors may vary among years.