Trait-based approaches can reveal the mechanisms through which disturbances or stress impact communities, allowing comparisons of the role of different mechanisms in shaping communities among taxonomic groups. Such information can lead to higher comparability, transferability and predictability of the outcome of restoration projects. However, multitaxa trait-based approaches were rarely used in the context of ecosystem restoration. We investigated the responses to environmental gradients of seven taxa (vascular plants, staphylinid and carabid beetles, spiders, isopods, diplopods and earthworms) in a restored floodplain using a species traits approach. We assessed the impact of flood disturbances and soil hydric stress on the functional diversity (FD) and community-weighted mean (CWM) response of traits for each taxon. Ordination of hydrological variables revealed two main gradients. The first was related to the spatiotemporal dynamics of flood disturbances and the second to the average changes in soil hydric conditions. The analysis of CWM revealed that larger, poorly mobile species with narrow ecological tolerances were filtered by regular floods and/or changes in soil hydric conditions. Functional diversity patterns differed between the two gradients: decreasing with increasing flood disturbance, but increasing along the soil hydric stress gradient. This suggests that the mechanisms shaping community composition differ between the two gradients with environmental filtering being dominant with increasing flood disturbances and competition decreasing with more soil hydric stress.Synthesis and applications. Our study shows that the impact of restored flood disturbances and soil hydric stress on plant and invertebrate functional diversity and community-weighted mean can be positive, negative or more complex depending on the taxonomic group and environmental gradient considered. The patterns can to some extent be explained by the specific characteristics of each group. Larger, poorly mobile species with narrow ecological tolerances were particularly vulnerable to changes in disturbance and stress regime following floodplain restoration. These species may therefore be lost in the initial phases of restoration projects, but other more characteristic species of dynamic floodplains will be favoured. Understanding the consequences of these contrasted responses for biodiversity conservation and ecosystem functioning constitutes the next challenge for ecosystem restoration. Our study shows that the impact of restored flood disturbances and soil hydric stress on plant and invertebrate functional diversity and community-weighted mean can be positive, negative or more complex depending on the taxonomic group and environmental gradient considered. The patterns can to some extent be explained by the specific characteristics of each group. Larger, poorly mobile species with narrow ecological tolerances were particularly vulnerable to changes in disturbance and stress regime following floodplain restoration. These species may therefore be lost in the initial phases of restoration projects, but other more characteristic species of dynamic floodplains will be favoured. Understanding the consequences of these contrasted responses for biodiversity conservation and ecosystem functioning constitutes the next challenge for ecosystem restoration.