Forest restoration treatments indirectly diversify pollination networks via floral- and temperature-mediated effects

<p>In North American conifer forests a variety of federally initiated thinning programs are implemented to restore pre-European settlement forest structures, but these changes may impact ecosystem function via impacts on sensitive biotic communities. Across the wildland-urban interface of the Front Range region of Colorado, agencies associated with the 'Collaborative Forest Landscape Restoration Program' (CFLRP) have implemented thinning treatments across thousands of hectares of ponderosa pine forest; here we leverage these treatments as an experimental framework to examine thinning effects on a pollinator community. We measured variation in forest structure and sampled bee community assemblages using multiple methods (trapping and netting) to compare bee biodiversity and patterns of floral visitation by bees (bee-flower networks) between mechanically thinned stands that were 3–10 years post-treatment and non-thinned stands. Three key findings emerged: (1) Native bee abundance, richness, and diversity were 120, 53, and 37% greater, respectively, in thinned stands. Nestedness, richness, and abundance of bee-flower interactions were all substantially higher in thinned stands, and there was increased functional redundancy in bee assemblages after thinning. (2) Structural equation modelling indicated that variation in temperature and floral abundance were mediated by canopy openness and were correlated with bee richness and abundance, thereby indirectly driving variation in bee-flower interactions. (3) Four floral species (<em>Penstemon virens, Cerastium arvense, Erysimum capitatum, </em>and<em> Geranium caespitosum</em>) were identified as key connectors in bee-flower interaction networks, though these were not necessarily the most abundant flowering plants. Our analyses indicate that native bee α-diversity and bee-flower interactions positively responded to thinning treatments, and these effects were indirectly driven by canopy removal. We conclude that CFLRP treatments have conservation value for native bee communities. Further monitoring is warranted to evaluate the longevity of these effects.</p>