Quantifying the roles of seed dispersal, filtering, and climate on regional patterns of grassland biodiversity.

Seed dispersal and local filtering interactively govern community membership and scale up to shape regional vegetation patterns, but data revealing how and why particular species are excluded from specific communities in nature are scarce. This lack of data is a missing link between our theoretical understanding of how diversity patterns can form and how they actually form in nature, and it hampers our ability to predict community responses to climate change. Here, we compare seed, seedling, and adult plant communities at 12 grassland sites with different climates in southern Norway to examine how community membership is interactively shaped by seed dispersal and local filtering, and how this process varies with climate across sites. To do this, we divide species at each site into two groups: locally transient species, which occur as seeds but are rare or absent as adults (i.e., they arrive but are filtered out), and locally persistent species, which occur consistently as adults in annual vegetation surveys. We then ask how and why locally transient species are disfavored during community assembly. Our results led to four main conclusions: (1) the total numbers of seeds and species that arrived, but failed to establish locally persistent populations, rose with temperature, indicating an increase in the realized effects of local filtering on community assembly, as well as an increase in the number of species poised to rapidly colonize those warmer sites if local conditions change in their favor, (2) locally transient species were selectively filtered out during seedling emergence, but not during seedling establishment, (3) selective filtering was partly driven by species climate preferences, exemplified by the poor performance of seeds dispersing outside of their realized climate niches into colder and drier foreign climates, and (4) locally transient species had traits that likely made them better dispersers (i.e., smaller seeds) but poorer competitors for light (i.e., shorter statures and less persistent clonal connections) than locally persistent species, potentially explaining why these species arrived to new sites but did not establish locally persistent adult populations. Our study is the first to combine seed, seedling, and adult survey data across sites to rigorously characterize how seed dispersal and local filtering govern community membership and shape climate-associated vegetation patterns.