PS 83-227
The effect of pollinator restorations on wild bee community composition
Recent concern over wild, native bee declines has prompted efforts to restore pollinator communities. The main restoration technique used is planting native wildflowers of pollinator-attractive species. Although several studies have shown that these pollinator restorations increase the species richness and abundance of native pollinators, less is known about how they affect pollinator community composition. We studied wild bee communities in pollinator restorations (size range 0.1 – 1 ha) at 15 sites throughout New Jersey and eastern Pennsylvania over three years. We paired each pollinator restoration with a nearby control situated in old-field habitat. Old fields, which are the naturally regenerating early/mid successional stage in our study region, support a high abundance and richness of native bees. Thus they serve as a conservative control with respect to the value of actively managed pollinator restoration. We compared wild bee species composition between pollinator restorations and old-field controls using a PERMANOVA. We then used a betadispersion test to determine whether species turnover across the 15 sites differed among pollinator restorations and old-field controls. Finally, we used indicator species analysis to evaluate whether certain bee species characterized either pollinator restorations or old-field controls.
Results/Conclusions
We collected a total of 5064 specimens of 136 bee species across all sites and years combined. While there was statistically significant effect of pollinator restoration on bee species composition, the size of the effect of small (R2 = 0.1; P<0.01) suggesting that in general, native bee communities are similar between pollinator restorations and old-field controls. The betadispersion test indicated that turnover was higher among old-field control sites than pollinator restorations (P<0.01). This indicates that old-field controls have more variable bee community composition among sites, as compared with pollinator restoration sites, which support more similar communities. Finally, seven bee species (Bombus bimaculatus, B. impatiens, B. griseocollis, Ceratina calcarata, Halictus ligatus, H. rubicundus, and Lasioglossum imatatum) characterized pollinator restorations as compared with controls. The greater compositional similarity among pollination restorations might be due to greater compositional similarity of the plant communities at restored sites, insofar as restorations use a subset of native plant species known to be attractive to pollinators. In combination with other work showing that overall, pollinator restorations increase wild bee abundance and richness our study indicates that while restorations benefit pollinator communities in a broad sense, they may confer the greatest benefits to a subset of species.