Comparison of grassland plant-pollinator networks on dairy farms in three contrasting French landscapes

• Published in: Acta oecologica (Montrouge) 2021-10, Vol.112, p.103763, Article 103763
• Main Authors: Michelot-Antalik, Alice, Michel, Nadia, Goulnik, Jérémie, Blanchetête, André, Delacroix, Emile, Faivre-Rampant, Patricia, Fiorelli, Jean-Louis, Galliot, Jean-Noël, Genoud, David, Lanore, Laurent, Le Clainche, Isabelle, Le Paslier, Marie Christine, Novak, Sandra, Odoux, Jean-François, Brunel, Dominique, Farruggia, Anne
• Format: Article
• Language: English
• Subjects: Agricultural practice; Diptera; DNA bar-coding; Ecology, environment; Hymenoptera; Life Sciences; Metagenomics; Pollination
• ISSN: 1146-609X; 1873-6238
• DOI: 10.1016/j.actao.2021.103763

Temperate grasslands provide both habitats and flower resources for pollinators in agricultural landscapes. Plant-pollinator networks change according to local and landscape variables, which are important to identify to help conserve pollinators in grasslands. We analysed plant-pollinator networks in 18 grasslands on experimental dairy farms located in three French regions contrasting by their climate, altitude, landscape or management. We combined visual surveys and pollen DNA barcoding. Our objectives were to determine which environmental factors influence pollinator taxa abundance and diversity and differences among the visual plant-pollinator networks in the three farming regions. Flower-visiting insects were trapped in six grasslands per farm during three sessions from mid-April to mid-July along fixed 400 m2 transects. Insects were identified individually to the lowest taxonomic rank possible. Pollen carried by insects was identified using nuclear ribosomal ITS2 sequences belonging to the NCBI nucleotide database. The size and diversity of plant-pollinator networks were much larger and higher in permanent grasslands at the two farms located in lowlands (Mirecourt) and mountains (Marcenat) than those at the farm with temporary grasslands and a crop landscape (Lusignan), but the degree of specialisation (H2′) was relatively similar and low (mean of 0.46). Diptera, especially Empididae and Syrphidae, represented most plant-pollinator interactions in Mirecourt and Marcenat, while Hymenoptera were more abundant at Lusignan. The percentage of semi-natural habitats in 500 m buffers and vegetation height explained 23% of the variance in pollinator abundance. Ranunculus sp. Knautia arvensis, Centaurea jacea and Trifolium repens were key plant species in the networks. DNA metabarcoding of pollen loads identified 114 genera in addition to those identified by visual observations (+34–42 per site), reflecting insects’ floral pathways and differences in the immediate landscape among farms. This study highlighted the importance of Diptera in plant-pollinator networks and the need to conserve permanent grassland diversity to conserve pollinators. •Size and diversity of plant-pollinator networks were contrasted between dairy farms.•Vegetation height and landscape composition influenced the pollinator abundance.•Diptera were the most abundant in permanent grasslands.•Metabarcoding supplemented the plant genera identified by visual networks.