Ecological restorations that re-establish plant communities are assumed to benefit animal communities, which may re-colonize restored areas from the surrounding landscape. Recent work shows that plant-based restorations are successful at establishing bee diversity and abundance at levels that approach remnant habitats. However, the community assembly process by which bee species from a regional pool are sorted into restored habitats during colonization is not well understood. In existing restorations, understanding the site-level habitat associations of bee species can shed light on the past community assembly process and inform restoration practice. Bee species are expected to occur in sites with characteristics that match up with bee functional traits. We used a multi-year dataset of bee specimens collected in a tallgrass prairie restoration to determine how bee species are distributed among a chronosequence of restored sites. Restored sites in our study area varied in their age (time since restoration) and several covariates: flowering plant diversity, recent burn history, presence of bison, and proximity to non-prairie habitats (agriculture, forest, and pasture). We expected bee functional traits including nesting substrate, sociality, lecticity, and phenology to predict species composition among restored sites.
We collected 5892 bees of 86 species in 2015 and 2016 at Nachusa Grasslands (Illinois). Our 14 sites compose a chronosequence of restored prairie 2-28 years in age with three remnant prairie controls and three agricultural (corn) controls. We used the fourth corner method to analyze the relationships between environmental site characteristics and species traits. This method quantifies environment-by-trait associations by combining three data matrices: sites by environmental characteristics, sites by species, and species by traits. Our preliminary analysis shows nonsignificant trends in bee-habitat associations, particularly with respect to sociality. Older restored sites and remnant sites may have more eusocial bees (e.g. Bombus (bumblebees) and Lasioglossum; p>0.1). Sites with prairie edges (compared to forest, agriculture, or pasture edges) may have more facultatively social bees (native bees in the genera Agapostemon, Ceratina, Halictus, and Lasioglossum; p>0.1). These trends inform restoration management by highlighting the importance of long-term restorations for social bees such as Bombus species, many of which are declining. Further, these trends suggest that well-connected prairie restorations may be particularly important for four diverse genera of native bees. Additional analysis in a regression framework will shed more light on the role of bee species traits in determining their habitat associations.