Linking macrodetritivore distribution to desiccation resistance in small forest fragments embedded in agricultural landscapes in Europe

International audience; Purpose Most of the agricultural landscape in Europe, and elsewhere, consists of mosaics with scattered fragments of semi-natural habitat like small forest fragments. Mutual interactions between forest fragments and agricultural areas influence ecosystem processes such as nutrient cycling, a process strongly mediated by the macrodetritivore community, which is however, poorly studied. We investigated macrodetritivore distribution patterns at local and landscape level and used a key functional trait (desiccation resistance) to gain mechanistic insights of the putative drivers. Methods Macrodetritivores were sampled in forest edges-centres of 224 European forest fragments across 14 landscapes opposing in land use intensity. We used a multilevel analysis of variance to assess the relative contribution of different spatial scales in explaining activity-density and Shannon-diversity of woodlice and millipedes, together with a model-based analysis of the multivariate activity-density data testing the effect on species composition. Secondly, we tested if desiccation resistance of macrodetritivores varied across communities at different spatial scales using linear mixed effect models. Results Forest edge-centre and landscape use intensity determined activity-density and community composition of macrodetritivores in forest fragments, while fragment characteristics like size and continuity were relatively unimportant. Forest edges and higher intensity landscapes supported higher activity-density of macrodetritivores and determined species composition. Forest edges sustained woodlouse communities dominated by more drought tolerant species. Conclusions Landscape use intensity and forest edges are main drivers in macrodetritivore distribution in forest fragments with desiccation resistance a good predictor of macrodetritivore distribution. Key functional traits can help us to predict changes in community structure in changing landscapes.