Environmental correlates of wild bee assemblages in pastoral landscapes in Ireland (Hymenoptera, Apidae)

Declines in wild bees have been reported internationally and attributed to habitat loss associated with agriculture. Recent research has led to a better understanding of the responses of bumblebees, especially within mixed, arable or fruit-growing agricultural landscapes. However gaps remain in scientific understanding of the responses of bees, particularly of solitary bees, to agricultural intensification in pastoral landscapes. In general, studies of bee declines have focussed on measures such as total abundance, species richness and diversity indices and there is a need for studies of changes in assemblages’ species composition in order to identify resilient and vulnerable species. The aim of this study is to examine the responses of solitary bees and bumblebees to grassland intensification in lowland, pastoral landscapes. The study identifies factors that influence wild bee diversity, abundance and assemblage composition. The relative importance of landscape composition; agricultural management (considered at the field, farm and landscape level); habitat structure and quality and more immutable environmental conditions such as latitude and longitude, altitude and edaphic factors on bee responses are studied. Based on differences in foraging distances and social behaviour, it is likely that solitary bees and bumblebees respond to anthropogenic change occurring at different scales. This hypothesis is tested. A field survey of bees across fifty agricultural sites, together with a survey of the environmental conditions at these sites was undertaken. Correlations were investigated using Mantel’s tests and Mantel correlograms, Procrustean rotations and indirect gradient analyses with ordinations. The relative importance of environmental variables was evaluated using a combination of methods. The variance in species composition of bee assemblages was decomposed between environmental predictors using distance-based Redundancy Analysis (RDA). Bayesian and Information theoretic methods were used to evaluate the relative importance of predictors of bee diversity and abundance. The abundance and species richness of bees in pastoral landscapes in Ireland are impacted by intensifying grassland management, with solitary bees showing a response to intensification at the field scale and bumblebees to intensification at the landscape scale. A shift in bee assemblage composition from assemblages dominated by diverse solitary bees to assemblages dominated by a few common bumblebees and a small number of solitary bee species was observed as field management became more intense. This gradient was also associated with a calcicole-calcifuge vegetation gradient and it is not possible to distinguish between the effects of the two factors on bee assemblage. Solitary bees showed greater sensitivity to site management than bumblebees, possibly due to previous local extinctions of more sensitive bumblebee species from the majority of farm sites. A model of extinction order from wild bee assemblages in the face of intensifying grassland management is proposed. There is high natural variability of bee assemblage composition, associated with solitary bees. Biogeographical influences were more influential in shaping assemblage composition than agricultural management. Spatial effects upon bees were significant for distances up to approximately 10km. Associations with vegetation contributed to this spatial pattern. Autogenic factors were also influential in spatial patterning of solitary bees. The mechanisms by which intensification of grassland management impacts on bees remain to be identified. To date, agri-environmental schemes have focused on the conservation of hedgerows. The conservation of bees in pastoral landscapes requires initiatives focused on grassland management at both the field and landscape scale. This will aid the conservation of common bee species and the maintenance of pollination services across regions. Rarer bees require conservation initiatives targeted at the locations where they survive. This study has suggested characteristics of such locations which may aid in their identification. The proposed ‘order of extinction’ model requires validation. It has potential to be used in quantifying the level of impact on bee assemblages and in monitoring the effects of environmental degradation and restoration on wild bees.