In order to understand the effect of disturbance on ecosystem functioning, one must take into account its indirect effects resulting from species interactions. While few empirical studies assess indirect effects and consider multiple trophic levels, recent statistical and theoretical advances provide new paths to do so. Here we investigate effects of disturbances in a boreal forest on multitrophic communities using a combination of the network and a trait-based approaches. Increasing functional diversity at different trophic levels is known to increase ecosystem functions, and disturbances may select for specific species traits. We therefore used structural equation modeling to quantify direct and indirect effects of environmental changes, on vegetation and soil fauna functional composition and on leaf litter decomposition. Forest disturbance intensity was found to have a bottom-up effect on species community composition, from lower trophic levels up to top predators. Disturbances also affected litter decomposition through changes of detritivore species composition. Such multitrophic assessments provide a more holistic understanding of ecosystem responses to environmental change.