Contrasting effects of wooded and herbaceous semi-natural habitats on supporting wild bee diversity

The use of pesticides and the lack of resources caused by the simplification of landscapes are often cited as the main drivers of the loss of wild bees. Landscape complexity has been shown to interact with local pesticide use in studies of the effect of these factors on the abundance and diversity of insects. But, to date, few studies have simultaneously investigated local and landscape effects on wild bees using more precise descriptors of the landscape and of farming practices. The aim of the present study was thus to disentangle the effects of landscape composition and farming practices on the taxonomic and functional diversity of wild bees and their possible consequences for pollination potential. We analysed a dataset of 107 crops at whose edges wild bees were captured between 2013 and 2019. Pesticide treatments of each crop were recorded throughout the year. At the landscape level, the proportions of different types of land use were calculated in a circle with a radius of 500 m. We caught 1536 wild bees belonging to 86 species and, using generalized mixed modelling, found that the number of insecticide applications had a negative effect on wild bee abundance and richness, while forest cover within a 500-m radius had a positive effect on wild bees whatever the farming practices. However, we also showed that the effect of the proportion of permanent grasslands on taxonomic and functional diversity of wild bees differed depending on the use of fungicides in the field. Finally, we found a negative effect of the number of herbicide applications on pollination potential. This study advances our understanding of the contrasting but complementary roles of diverse semi-natural habitats in agricultural landscapes in supporting wild bee diversity and which, under certain conditions, may buffer the detrimental consequences of pesticide use.