SCALE-DEPENDENT EFFECTS OF LANDSCAPE CONTEXT ON THREE POLLINATOR GUILDS

Most ecological processes and interactions depend on scales much larger than a single habitat, and therefore it is important to link spatial patterns and ecological processes at a landscape scale. Here, we analyzed the effects of landscape context on the distribution of bees (Hymenoptera: Apoidea) at multiple spatial scales with respect to the following hypotheses: (1) Local abundance and diversity of bees increase with increasing proportion of the surrounding seminatural habitats. (2) Solitary wild bees, bumble bees, and honey bees respond to landscape context at different spatial scales. We selected 15 landscape sectors and determined the percentage of seminatural habitats and the diversity of habitat types at eight spatial scales (radius 250–3000 m) by field inspections and analyses of vegetation maps using two Geographic Information Systems. The percentage of seminatural habitats varied between 1.4% and 28%. In the center of each landscape sector a patch of potted flowering plants (four perennial and two annual species) was placed in the same habitat type, a grassy field margin adjacent to cereal fields. In all, 865 wild bee individuals and 467 honey bees were observed and an additional 475 individuals were caught for species identification. Species richness and abundance of solitary wild bees showed a close positive correlation with the percentage of seminatural habitats at small scales up to 750 m, whereas bumble bees and honey bees did not respond to landscape context at these scales. In contrast, honey bees were correlated with landscape context at large scales. The densities of flower-visiting honey bees even increased with decreasing proportion of seminatural habitats at a radius of 3000 m. We are not aware of any empirical studies showing contrasting foraging patterns related to landscape context at different spatial scales. We conclude (1) that local landscape destruction affects solitary wild bees more than social bees, possibly changing mutualistic plant–pollinator and competitive wild bees–honey bees interactions and (2) that only analyses of multiple spatial scales may detect the importance of the landscape context for local pollinator communities.