Multiscale landscape drivers of contemporary pollen flow in the insect-pollinated herb, Pulsatilla vulgaris

Understanding the drivers and spatial scale of gene flow is essential for the management of species living in fragmented landscapes. In plants, contemporary pollen flow is commonly modeled as a functional of inter-mate distance, with pollen dispersal declining exponentially within a short distance of mother plants. However, parentage-based reconstruction of realized pollen flow has uncovered pervasive dispersal over surprisingly large distances for animal-pollinated species 1,2. This suggests that geographic distance alone explains only a fraction of the inter-individual and inter-population variation in mating, and indeed often fails to sufficiently capture variation at the tail end of dispersal kernels (i.e. long-distance dispersal events)3,4. Thus there is a critical need for researchers to take a more pollinator-eyed view of contemporary gene flow by pollen in plants by (1) considering additional factors that may alter attractiveness, detectability, and accessibility of plants to pollinators beyond the effects of distance (e.g. landscape context and landscape resistance), and (2) recognizing that within and among-population gene flow by pollen may depend on non-overlapping sets of pollinators that respond to these features at different spatial scales.

Here we quantify the effects of floral resources and landscape composition on contemporary pollen flow in the insect-pollinated herb, Pulsatilla vulgaris over multiple spatial scales. We reconstruct realized pollen flow and pollen immigration for seven fragmented populations in an agricultural matrix, and tested the hypothesis that within-population mating outcomes are related to resources and landscape context measured locally, and that among-population pollen flow is related to features measured at larger spatial scales. We found that within-population pollen flow was explained by features of individual plants and habitat patches; mean pollination distances and the proportion of selfed seeds per mother decreased as a function of floral density around maternal plants, and larger populations had significantly lower selfing rates. In contrast, among-population contemporary pollen flow (interpreted as pollen immigration rates) was related to features measured at the level of habitat patches and surrounding landscape; pollen immigration significantly increased with population size and the amount of seminatural landscape surrounding populations measured at a large spatial scale (1000 m radius), and significantly decreased with the amount of forest measured at an intermediate spatial scale (250 m radius). Together our results suggest that within- and among-population contemporary pollen flow may be governed by different underlying processes, possibly related to differences in the foraging range and landscape use of bee species that contribute to pollination at different scales.