Results
We caught in total 1419 individuals of 79 wild bee species, excludingApis mellifera (honey bee). Most bee individuals were caught in May (n=611) followed by June (n= 498) and August (n=310). Hoverflies were predominantly (92%) caught in August with 214 individuals of 21 species in total.
We observed 121 flowering animal-pollinated plant species (Table S1, Online Appendix 1). Flowering plant species richness ranged from one to 22 species near the pan traps with the highest number in June (mean±SD =11.55±5.4) followed by August (9.05±4.75) and May (7.41±3.45).
We found contrasting effects of arable field cover and landscape heterogeneity on both pollinator guilds (Fig. 2). Arable field cover negatively affected wild bee species richness. This effect was statistically clear (p<0.05) at small to intermediate spatial scales (140-400m) and peaked around 200m. Hence, dry grasslands that feature high proportion of arable fields in the surrounding show on average less wild bee species. In contrast, hoverfly species richness and abundance were positively affected by arable field cover at much larger scales (500-3000m), supporting hypothesis 1.
Landscape heterogeneity positively affected wild bees (Fig. 2, 3). Similar to the effect of arable field cover, the scale of effect for species richness had a peak at intermediate spatial scales (580m). Hoverfly species richness and abundance were negatively affected from intermediate to large spatial scales (~500m – 3000m). On these scales arable field cover and landescape heterogeneity were negatively related to eachother (Supplementary information, Appendix 2). At small spatial scales (~120m), landscape heterogeneity had a statistically clear positive effect on hoverfly abundance, indicating that a heterogeneous environment in the vicinity of dry grasslands increase the abundance of hoverflies. Overall, we found no support that landscape heterogeneity particularly enhance hoverflies compared to wild bees (hypothesis 2). Yet, our scale-crossing analyses showed that wild bees were mostly affected on smaller spatial scale compared to hoverflies (with the exception of the positive effect of landscape heterogeneity on abundance) supporting hypothesis 3.
We predicted that functional flower diversity in particular positively affects wild bees and flower traits associated with ‘attractiveness’ has a positive effect on hoverflies (hypothesis 4). Species richness of both guilds was not affect by any parameter of the plant community, except a negative effect of FDnectar access on hoverflies (Fig. 4). Abundance of wild bees and hoverflies was positively affected solely by CWMflower height. However, we observed also negative effects of plant community attributes on wild bee abundance (FDflower height, FDnectar access, CWMcolor yellow) and hoverfly abundance (FDnectar access). In summary, we found no evidence that functional diversity enhance pollinators, but the effect of CWMflower height, indicate that plant communities with conspicious flowers may attract pollinators.