Trait-based food web model reveals the underlying mechanisms of biodiversity-ecosystem functioning relationships.

The concept of biodiversity-ecosystem functioning (BEF) has been studied over the last three decades using experiments, theoretical models and more recently observational data. While theoretical models revealed that species richness is the best metric summarizing ecosystem functioning, it is clear that ecosystem function is explained by other variables besides species richness. Additionally, theoretical models rarely focus on more than one ecosystem function, limiting ecosystem functioning to biomass or production. There is a lack of theoretical background to verify how other components of biodiversity and species interactions support ecosystem functioning. Here, using simulations from a food web model based on a community assembly process and a trait-based approach, we test how species biodiversity, food web structure and predator-prey interactions determine several ecosystem functions (biomass, metabolism, production and productivity). Our results demonstrate that the relationship between species richness and ecosystem functioning depends on the type of ecosystem function considered and the importance of diversity and food web structure differs across functions. Particularly, we show that dominance plays a major role in determining the level of biomass, and it is at least as important as the number of species. We find that dominance occurs in the food web when species do not experience strong predation. By manipulating the structure of the food web, we show that species using a wider trait space (generalist communities) result in more connected food webs and generally reach the same level of functioning with less species. The model shows the importance of generalist versus specialist communities on BEF relationships, and as such, empirical studies should focus on quantifying the importance of diet/habitat use on ecosystem functioning. Our study provides a better understanding of BEF underlying mechanisms and generates research hypotheses that can be considered and tested in observational studies. We recommend that studies investigating links between biodiversity and ecosystem functions should include metrics of dominance, species composition, trophic structure and possibly environmental trait space. We also advise that more effort should be made into calculating several ecosystem functions and properties with data from natural multitrophic systems.