PS 24-30
Seeds, space and water: Assessing the relative contributions of spatial and environmental factors on seedbank and vegetation communities

Tuesday, August 11, 2015
Exhibit Hall, Baltimore Convention Center
Dylan E. Chapple, Environmental Science, Policy and Management, UC Berkeley, Berkeley,, CA
Mark Hammond, Kellog Biological Station, Michigan State Univeristy
Katherine L. Gross, Kellogg Biological Station, Michigan State University, Hickory Corners, MI
Katharine N. Suding, Ecology and Evolutionary Biology, University of Colorado, Boulder, CO

           Seeds play a vital role in structuring plant communities.  Seed arrival and establishment can have long lasting impacts on ecosystems and are mediated by the interaction between stochastic and deterministic factors.  The influence of these factors on community development is a central theme in ecology.  While much work with seeds has occurred in the context of manipulative studies, observational studies that incorporate both seed and plant cover data are less common.  Here, we contrast plant vegetation cover data taken across sand prairie/savannah habitats with the corresponding seedbank at each plot.  These sand prairies are discrete habitat patches surrounded by forest ecosystems inhospitable to their characteristic species, making them good habitats to explore how seed dispersal interacts with space and environmental factors.  We propose the following hypotheses to guide our study: 1) Both environmental filtering and dispersal (geographic distance) will influence patterns of vegetative and seedbank community composition 2) Environmental factors will play a stronger role in structuring vegetative communities, while seedbank composition will have a stronger spatial signature.


We find high dissimilarity between cover and seedbank using both Bray-Curtis (abundance-based) and Jaccard (presence/absence) dissimilarity indices.  Using variance partitioning, we find that of 4 environmental factors tested, soil water was the best model fit, explaining similar levels of variability in vegetation composition (34%) and seed community composition (35%).  Geographic distance explained additional variation that was unrelated to environmental variables in our dataset for both the seed (mantel statistic=0.15, p=0.0017) and vegetation (mantel statistic=0.13, p=0.0015) composition. Thus, despite the finding that the seed and vegetation were compositionally very different from one another, the importance of environment filters and geographic distance were surprisingly similar across the two types of communities. This result suggests that relative roles of stochastic and deterministic factors may be similar in structuring species represented in the seed bank and in the vegetative plant community.