COS 314-4 - Diversity-Interactions models for predicting the total response and breakdown by species proportions in biodiversity experiments

Abstract: Diversity-Interactions modelling is a regression-based approach for modelling the biodiversity and ecosystem function (BEF) relationship. It assumes that the main driver behind changes in ecosystem functions is the initial proportion of each species in the ecosystem and it quantifies how changing the composition, richness, and proportions of species in a community will affect ecosystem function. In a grassland experiment carried out in Ireland, species diversity was manipulated across 43 plots, with monocultures of six species (two grasses, legumes, and herbs) and 13 unique mixtures from two up to six species that systematically varied in species proportions. Earlier research on this study showed that both functional group richness and changes in sown species proportions had a strong effect on the biomass produced over a two-year period. However, it is also of interest to know how the realised species proportions of the biomass changed from the sown proportions over the two-year period and if it is possible to predict changes in species proportions over time by the sown proportions.

Here, we extend the Diversity-Interactions modelling approach to address how initial proportions jointly affect both the biomass and the proportional contributions to the biomass of the sown and invasive species over time. Applying our new approach to the Irish grassland data showed that sown species proportions were strongly related to the realised species proportions in the biomass, as well as the total biomass. We found that combining grass and legume species increased the proportion of total biomass achieved by grasses and that the inclusion of herbs in mixtures had the potential to reduce the proportion of unplanted species in the biomass.

Finally, we compare our extended Diversity-Interactions modelling approach to the additive partitioning method for the estimation of the selection and complementarity effects and assess its potential for providing insight into the biological mechanisms underlying BEF relationships.