Driving factors and temporal fluctuation of Collembola communities and reproductive mode across forest types and regions.

Despite the major role of Collembola in forest soil animal food webs, ecological and evolutionary determinants of their community composition are not well understood. We investigated abundance, community structure, life forms, and reproductive mode of Collembola in four different forest types (coniferous, young managed beech, old managed beech, and unmanaged beech forests) representing different management intensities. Forest types were replicated within three regions across Germany: the Schorfheide-Chorin, the Hainich, and the Swabian Alb, differing in geology, altitude, and climate. To account for temporal variation, samples were taken twice with an interval of 3 years. To identify driving factors of Collembola community structure, we applied structural equation modeling, including an index of forest management intensity, abiotic and biotic factors such as pH, C-to-N ratio of leaf litter, microbial biomass, and fungal-to-bacterial ratio. Collembola abundance, biomass, and community composition differed markedly between years, with most pronounced differences in the Schorfheide, the region with the harshest climatic conditions. There, temporal fluctuations of parthenogenetic Collembola were significantly higher than in the other regions. In the year with the more favorable conditions, parthenogenetic species flourished, with their abundance depending mainly on abiotic, density-independent factors. This is in line with the "Structured Resource Theory of Sexual Reproduction," stating that parthenogenetic species are favored if density-independent factors, such as desiccation, frost or flooding, prevail. In contrast, sexual species in the same year were mainly influenced by resource quality-related factors such as the fungal-to-bacterial ratio and the C-to-N ratio of leaf litter. The influence of forest management intensity on abundances was low, indicating that disturbance through forest management plays a minor role. Accordingly, differences in community composition were more pronounced between regions than between different forest types, pointing to the importance of regional factors.