Species-specific responses to landscape composition of wild bees in pear orchards

Pollinator-dependent fruit crops are widely grown in Xinjiang, China. However, the communities of wild bees that might contribute to crop pollination in these fruit-growing landscapes have not been well characterized, and it is not known how these communities are affected by landscape composition. We assessed the wild bee communities in 52 Korla fragrant pear (Pyrus sinkiangensis) orchards using colored pan traps from 2018 to 2021 and quantified how wild bee communities are influenced by landscape composition at four spatial scales (0.5, 1.0, 1.5, and 2.0 km). We also assessed wild bee abundance on ten wild plant species in the field margins of pear orchards using sweep netting after pear flowering in 2020 and 2021 to assess their possible function in supporting wild bees. A total of 3576 individuals of 28 wild bee species were collected in pan traps. Three species made up 90% of all individuals: Andrena yamagishi (57%), Lasioglossum pseudannulipes (17%), and Lasioglossum niveocinctum (16%). The abundance of A. yamagishi was positively associated with semi-natural habitat at 0.5–2.0 km scales while the abundance of L. pseudannulipes was positively associated with annual crop habitats at the same scales. There was no significant response of L. niveocinctum abundance to landscape composition. Neither total wild bee richness nor diversity was associated with landscape composition. Medicago sativa and Apocynum venetum in orchard margins had relatively high wild bee abundances. Our study shows that bee species within the same genus respond differently to landscape composition, highlighting the need for addressing landscape effects on pollinators at the species level rather than for the community as a whole.