Disentangling the contributions of dispersal limitation, ecological drift, and ecological filtering to wild bee community assembly

Abstract The species composition within communities is highly dependent on the rate of species immigration and whether immigrating species possess the functional traits required by the prevailing environmental conditions. Once established, random fluctuations in birth and death rates may reduce the diversity of ecologically equivalent species if local populations are not replenished by immigrating individuals. Consequently, three key processes drive community assembly: dispersal limitation, ecological filtering, and ecological drift. However, disentangling the relative contribution of these processes remains a challenge in community ecology. We used a binomial generalized linear mixed model to test whether the occurrences of solitary bees within 46 communities in southeast Norway were driven by (1) dispersal limitation, that is, the geographic distance to the nearest site where conspecifics occurred; (2) ecological filtering, that is, if forb species richness selected for non‐Ericaceae‐affiliated species; and (3) ecological drift, that is, if small, isolated communities were dominated by regionally common species. The regression slopes from the model for each potential driver of community composition were compared with those expected under a null model, in which species were treated as ecologically equivalent. Both dispersal limitation and ecological filtering influenced the probability of species occurring within communities. The occurrence of species decreased with elevation, and this relationship depended on the relative commonness of species and their floral preferences. For non‐Ericaceae‐affiliated species, the patterns of occurrence mirrored that expected under the null (neutral) model, resulting in the same patterns as would be expected under ecological drift. In contrast, the response of Ericaceae‐affiliated species differed from what would be expected from the null model. Our results also indicate that processes leading to neutral dynamics in species compositions drive a large part of the gradient in species richness in Norwegian bee communities. These processes seem related to sampling effects so that large and interconnected communities have a higher probability of including regionally rare species than small, isolated communities. Our results suggest that targeting habitats—where the influence of ecological filtering is expected to be greater than that of neutral dynamics—can increase the success of habitat management plans aimed at promoting rare species.