The strength of biodiversity–ecosystem functioning (BEF) relationships varies within and across studies, depending on the investigated ecosystem function and diversity facet (e.g., species richness or functional composition), limiting our ability to translate BEF results into recommendations for management and conservation. The variability in BEF relationships is particularly high when considering complex multitrophic communities and can be explained by food web contexts. Here we examine how different plant diversity facets affect biomass stocks and energy flows of each trophic group depending on their position in the trophic network. We used coupled aboveground–belowground multitrophic networks of energy dynamics, assembled across the experimental gradients of grassland plant species richness, functional diversity, and presence of plant functional groups. We compared the strengths of these diversity effects between trophic groups, trophic levels, aboveground versus belowground subnetworks, and types of ecosystem functions. Plant species richness, functional trait diversity, and the presence of legumes and grasses were influential drivers of ecosystem energetics. The effects of plant species richness across the food web often operated through mechanisms of plant functional-trait diversity. The effects of plant species richness attenuated across trophic levels. Legume presence strengthened the top-down control (predation) of primary consumers. We found an overall mismatch in the strength of diversity effects on flows versus stocks. Some trophic groups showed even contrasting direction in responses of their stocks and flows to plant diversity. This indicates that plant diversity constrains consumer functioning by means other than only altered consumer biomass. Responses of flows and stocks to plant diversity differed between trophic groups, and aboveground versus belowground parts. Individual stocks and energy flows were responsive to different biodiversity facets, highlighting the importance of the explicit consideration of individual functions and diversity facets for a comprehensive multitrophic understanding. For example, legume presence increased aboveground processes but reduced plant carbon uptake and belowground plant production. Plant communities containing legumes lost more biomass to herbivores, had faster decomposition, and channeled less energy to soil detritus. An important implication of these results is that targeted grassland management would profit from focusing on specific plant diversity facets depending on the ecosystem function or service of interest.
