Pollen dispersal patterns differ among sites for a wind-pollinated species and an insect-pollinated species

Tree cover and palm population structure determine patterns of palm-pollinator interaction networks in a grassland-forest ecotone

Pollen dispersal by insects determines the patterns of reproductive encounters between plants with flowers that have spatially or temporally segregated sexes. Pollinators exhibit varied responses to environmental gradients, such as those in grassland-forest ecotones. Individual-based interaction networks are useful yet underexplored tools to understand how interactions vary across these gradients. To test how a grassland-forest ecotone gradient affects these interactions, we studied pollination networks of Butia odorata individuals, a key palm tree species in a threatened South American grassland ecosystem. We assessed, through Structural Equation Models (SEMs), how network metrics (specialization and modularity) and pollinator richness and abundance (total, peripheral, core) respond to gradients of habitat and population structure in the grassland-forest ecotone. Networks with more spatially isolated palms showed greater specialization and modularity. Pollinator richness was dependent on the habitat context and pollinator role. The peripheral pollinators were negatively affected by palm density. The core pollinators were positively affected by tree cover, which, in turn, was positively associated with palm density and proximity to the forest. Our results suggest that palm population density can modulate the presence of peripheral pollinators, while increased tree cover in the grassland matrix can promote pollinator diversity by decreasing the dominance of core species.

Long-Distance Pollen Dispersal in Urban Green Roof and Ground-Level Habitats

Long-distance pollen dispersal is critical for gene flow in plant populations, yet pollen dispersal patterns in urban habitats such as green roofs have not been extensively studied. Pollen dispersal patterns typically are assessed either by fitting non-linear models to the relationship between the degree of pollen dispersal and distance to the pollen source (i.e., curve fitting), or by fitting probability density functions (PDFs) to pollen dispersal probability histograms (i.e., PDF fitting). Studies using curve fitting typically report exponential decay patterns in pollen dispersal. However, PDF fitting typically produces more fat-tailed distributions, suggesting the exponential decay may not be the best fitting model. Because the two approaches may yield conflicting results, we used both approaches to examine pollen dispersal patterns in the wind-pollinated Amaranthus tuberculatus and the insect-pollinated Solanum lycopersicum at two green roof and two ground-level sites in the New York (NY, United States) metropolitan area. For the curve fitting analyses, the exponential decay and inverse power curves provided good fits to pollen dispersal patterns across both green roof and ground-level sites for both species. Similar patterns were observed with the PDF fitting analyses, where the exponential or inverse Gaussian were the top PDF at most sites for both species. While the curve fitting results are consistent with other studies, the results differ from most studies using PDF fitting, where long-distance pollen dispersal is more common than we observed. These results highlight the need for further research to compare curve and PDF fitting for predicting pollen dispersal patterns. And, critically, while long-distance pollen dispersal may be an important component of overall pollen dispersal for A. tuberculatus and S. lycopersicum in both urban green roof and ground-level sites, our results suggest it potentially may occur to a lesser extent compared with plants in less-urban areas.