COS 158-4 - Bee-wary? Modeling the influence of honey bees (Apis mellifera) on wild pollinator occupancy and detection in Mid-Atlantic (U.S.A.) gardens

Background/Question/Methods

The nearly worldwide expansion of honey bees (A. mellifera), especially outside of its native range, has greatly impacted plant-pollinator and ecosystem interactions from historical baselines. A. mellifera and other managed beesrequire locally-foraged resources to meet their high energy demands, often at the expense of wild pollinators. This is especially true in urban ecosystems, where the potential for resource competition between A. mellifera and other pollinators is an important consideration for conservation efforts. In this study, our aim was to assess the impact A. mellifera had on the presence of wild, predominantly native pollinators throughout a Mid-Atlantic (U.S.A.) region consisting of sites along an urban to rural gradient. We accomplished this by visually monitoring pollinator activity on Dense Blazing Star (Liatris spicata; Asteraceae) and Clustered Mountain Mint (Pycnanthemum muticum; Lamiaceae), two commonly cultivated species native to the region. We then analyzed the impact of A. mellifera co-occurrence on the probability of occupancy (p) and detection (psi) of other pollinators using the Program Presence.

Results/Conclusions

We observed a diverse range of pollinator occupants across all sites. The bee occupants were predominantly Xylocopa virginica or A. mellifera species, from the Bombus genera, or Halictidae family members. On L. spicata, Bombus spp. were the most common at 32.4%. The occupancy and detection of all pollinators on L. spicata inflorescences was negatively correlated with the occupancy of A. mellifera (Beta = -40.08). Comparing bees, A. mellifera occupancy negatively impacted Bombus spp. (Beta = -2.76) and X. virginiana (Beta = -0.38) detection, whereas Halictidae family bees were not significantly affected.


A. mellifera was the dominant pollinator species on P. muticum, comprising 42.6% of the individuals identified; however, A. mellifera more often co-occurred with wild pollinators. When modeling all P. muticm pollinators, A. mellifera negatively impacts detection (Beta = -0.58) but the null model was also supported (AIC weight = 0.29). The detection of Bombus spp., X. virginica, Halictidae bees, or wasps was not significantly affected. These results suggest that P. muticum best supports a higher diversity of wild urban pollinators, providing an abundance of resources even when managed bees are present. Alternatively, competition for L. spicata resources favors A. mellifera.