Effects of land use on bee diversity

Abstract

Bees are a diverse group of insects specialised to use floral resources throughout their lifecycle, making them one of the most important groups of pollinators. Land-use change and agricultural intensification can reduce the diversity of these species, with potential consequences for plant communities and a number of globally-important crops. Some key knowledge gaps limit our ability to inform conservation activities, however. Which species or groups of species are most vulnerable to human land use and agricultural intensification? Do different aspects of bee community diversity respond similarly to land-use pressures? Are the same patterns repeated around the world? Where has bee diversity been most heavily impacted? I have collated a global dataset of bee diversity from the published literature and integrated it with information on land use, species' traits and phylogeny in order to answer these questions. I have shown that functional traits interact in complex ways with land-use pressures in order to influence the occurrence and abundance of European bees; in spite of these complexities, general lessons can still be learnt about management priorities. I have found that species sensitivities to some---but not all---land-use transitions are phylogenetically non-random, highlighting particularly vulnerable clades. These functional and phylogenetic differences scale up to influence community-level diversity, but there are spatial mismatches: taxonomic diversity responds differently to land-use pressures than do functional and phylogenetic diversity. While the dataset I have collated is the most extensive of its kind, the majority of available, accessible data are from European and North American bumblebees - a bias that has also affected previous similar analyses. I show that these biases can affect predictions of land-use impacts on bee diversity if regional differences in diversity, biotic sensitivity and threat intensity are not considered, highlighting the importance of primary data collection in undersampled regions, which are highly economically vulnerable to pollination deficits.