The effect of manipulated plant diversity on invertebrate functional groups of different trophic levels in an experimental grassland

Background/Question/Methods Based on data collected in the Jena Experiment, we used multivariate statistical techniques and constructed interaction webs to study the relationship between plant species richness (S), number and species richness of plant functional groups (FG), presence/absence of specific plant FGs, and plant biomass as predictor variables and as response variables 6 herbivore, 9 zoophagan and 3 decomposer FGs. The consumer FGs comprised 887 insect and spider species with a total of about 80,000 individuals. We also took into account the relationship between the consumer FGs and an additional 23 abiotic and biotic response variables (Nonmetric Multidimensional Scaling axes scores) extracted from 487 original variables measured by various other project teams.

Results/Conclusions

While S, population density (N) and Shannon diversity (H) of herbivores were predominantly positively correlated to S and the number of FGs of plants, the effect of plant biomass was variable. Presence or absence of particular plant FGs (e.g. grasses or legumes) had a strong effect on some herbivore FGs. Zoophagans showed the same positive response except for ground-dwelling carabid beetles which were negatively affected, in particular by plant biomass. Of the decomposer groups, only phytosaprohagous dipterans were strongly positively correlated with N, S and H of plants. Both S and the number of FG of plants had a strongly positive effect on the number and strength of interactions in interaction webs comprising three trophic levels (plants, herbivores and predators/parasitoids). This showes that reduced plant species and functional diversity may have negative consequences for the functioning of ecosystems.