Contrasting effects of grassland management modes on species-abundance distributions of multiple groups

Intensive land use is a major cause of biodiversity loss, but most studies comparing the response of multiple taxa rely on simple diversity measures while analyses of other community attributes are only recently gaining attention. Species-abundance distributions (SADs) are a community attribute that can be used to study changes in the overall abundance structure of species groups, and whether these changes are driven by abundant or rare species. We evaluated the effect of grassland management intensity for three land-use modes (fertilization, mowing, grazing) and their combination on species richness and SADs for three belowground (arbuscular mycorrhizal fungi, prokaryotes and insect larvae) and seven aboveground groups (vascular plants, bryophytes and lichens; arthropod herbivores; arthropod pollinators; bats and birds). Three descriptors of SADs were evaluated: general shape (abundance decay rate), proportion of rare species (rarity) and proportional abundance of the commonest species (dominance). Across groups, taxonomic richness was largely unaffected by land-use intensity and only decreased with increasing mowing intensity. Of the three SAD descriptors, abundance decay rate became steeper with increasing combined land-use intensity across groups. This reflected a decrease in rarity among plants, herbivores and vertebrates. Effects of fertilization on the three descriptors were similar to the combined land-use intensity effects. Mowing intensity only affected the SAD descriptors of insect larvae and vertebrates, while grazing intensity produced a range of effects on different descriptors in distinct groups. Overall, belowground groups had more even abundance distributions than aboveground groups. Strong differences among aboveground groups and between above- and belowground groups indicate that no single taxonomic group can serve as an indicator for effects in other groups. In the past, the use of SADs has been hampered by concerns over theoretical models underlying specific forms of SADs. Our study shows that SAD descriptors that are not connected to a particular model are suitable to assess the effect of land use on community structure.