SD28: Forest bees need compositionally diverse canopies

Forests provide critical habitats for pollinating insects, including forest-dependent and habitat generalist species, yet it is unknown how these assemblages are shaped by overstory tree composition. We sampled bees in closed canopy plots in the southeastern United States representing a continuum of forest age and tree composition, from conifer-dominated forests to forests dominated by deciduous broadleaf trees. Species-specific responses of bees to forest composition, and the influence of their traits on responses, were estimated using a joint species distribution model. Additionally, we investigated species richness trends of nesting, sociality, and phenological trait groups. Forest composition greatly influenced bee species occurrence: 48% of species had positive relationships with the diversity of insect-pollinated broadleaf trees and 46% had negative relationships with the proportion of conifer basal area. Ground nesting bees and their parasites with early phenological activity (e.g., Andrena, Colletes, Eucera, Hylaeus, and Nomada) drove these responses and richness patterns supported these trends. Ground nesting bees active later in the year (e.g., most Lasioglossum) were generally unresponsive to forest composition. Our results indicate that phenology is an important factor determining bee species’ forest dependency and sensitivity to forest composition. We conclude that diverse broadleaf forests are crucial to maintaining bee diversity by providing floral resources that support forest-dependent species even in closed-canopy conditions. Although open-canopy conifer forests provide valuable habitat for bees, a lack of traits indicating dependency on conifer forests or trees suggests these management practices may more strongly favor habitat generalists than forest specialists.