Changes in grassland plant-pollinator network structure with fire and grazing

Pollinating insects are key community members of grassland systems, providing the ecosystem service of pollination of flowering plants. Recent work in the field of ecological networks has emphasized the importance of understanding the structure of interactions between plants and pollinators at the community scale for system robustness to species loss. Several hypotheses exist as to which network structure or arrangement of species interactions is optimal for system robustness. Proposed network structures include nestedness, decreasing specialization of interactions, and a modular arrangement. Extinction cascade models, or simulations of species removal, can also provide estimates of system robustness. Here we examine changes in network structure and robustness of plant pollinator networks across Konza Prairie Biological Station, a tallgrass prairie in northeast Kansas and landscape-scale experiment subject to different fire frequency and ungulate grazing treatments. Fire and grazing are the major disturbance regimes characteristic of grassland systems. Using stepwise selection of a partial redundancy analysis, presence or absence of large ungulate grazers, historic fire frequency and sampling date were found to be the most important variables explaining variation in network structural properties. The final model explained 30% of the variation in plant-pollinator network structure with grazers being the most important variable, explaining 20.9% of the variation. Network nestedness was negatively correlated with presence of grazers (t=-2.81, n=24, P=0.01). Specialization was negatively correlated with historic burn interval (R²=0.17, F_1,22=4.45, P=0.04). Our results show that grassland plant-pollinator network structure does change with fire and grazing regimes and these effects should be considered when managing for pollinator conservation.