Rare fish species in European lakes – patterns and processes

According to differences in life history and species niches, community assembly processes are predicted to differ between common and rare species. While neutral processes, in particular dispersal, should contribute strongly to assembly of rare species, environmental filtering (species sorting) should dominate the community assembly of common species. We analysed commonness and rarity by occurrence, abundance and geographic range among 82 fish species in 1871 European lakes and reservoirs. The common 25% (21 species) were excluded, and structure and assembly processes of the 75% rare (61 species) species occurring in 348 lakes were evaluated. General linear latent variable models indicated that environmental variables related to lake type (lakes vs. reservoirs), climate and ecosystem size predicted a large proportion of variance for both rare species presence/absence and abundance models, while spatial variables (co-occurrence in watersheds) contributed little to the models. To link community structure with assembly processes, the contribution of seven fish traits related to life history, feeding and habitat preference to predicting the geographic range of the rare species was analysed by boosted regression trees. Intermediate average fecundity and high maximum body length of species predicted smaller geographic range and hence a higher level of rarity, but the response curves of the variables were mostly non-linear and difficult to interpret. In contrast to our assumption, the results of the dominant assembly processes and predictors for rare species were very similar to those predicted for the common fish species in lakes at the continental scale.