LB 33-280 - Grassland restoration mediates plant-pollinator species interaction network reassembly

The maintenance of species-level diversity is a primary goal of grassland restoration, but it is often complicated by the potentially distinct influences of common restoration practices, such as prescribed fire, on target taxonomic groups. Further, beyond species composition, it is essential for effective restoration to consider the impacts of management on species interaction networks, especially for mutualistic species interactions, such those between plants and pollinators. In this project, we quantified the impact of three commonly employed grassland restoration practices on plant-pollinator network composition. Specifically, we evaluated the effect of 1) large-scale prescribed burning, 2) small-scale prescribed burning, and 3) small-scale burning plus seeding, relative to control plots, in each of 10 geographically separated grassland systems in the Cross Timbers ecoregion of Texas and southern Oklahoma. We compared pre- and post-restoration grassland plant-pollinator species interaction networks to determine the extent to which plant and pollinator communities and their species interactions reassemble following a restoration intervention, and to additionally identify if key local and landscape factors, such as local bare ground availability and surrounding grassland habitat cover impacted network architecture. Restoration-induced habitat architecture changes, especially bare ground cover, were found to have significant impacts on plant-pollinator interaction networks. Across all treatments, burning increased bare ground cover, which was positively correlated with network robustness and overall pollinator abundance. In other words, sites with more bare ground cover had greater numbers of pollinators and greater redundancy in network architecture. Likewise, bare ground cover was negatively correlated with network specialization, indicating that plant-pollinator interactions were more generalist in sites with more bare ground, including those sites that recently received the prescribed burn treatments. These results indicate that restoration activities might increase network redundancy and reduce overall levels of specialization within mutualistic interaction networks. Further, local ground cover management may play a critical role in maintaining these complex species-interaction networks. Future research will further investigate the impact of restoration treatment on different plant and pollinator functional groups and differences in the interaction network architecture over multiple years post-restoration.