OOS 26-3
Quantifying the interaction between stochastic and deterministic processes in microbial community assembly: Results from a trait-based model

Wednesday, August 13, 2014: 2:10 PM
202, Sacramento Convention Center
Sarah E. Evans, Kellogg Biological Station, Michigan State University, Hickory Corners, MI
Steven D. Allison, Ecology and Evolutionary Biology/Earth System Science, University of California, Irvine, CA
Background/Question/Methods

Community composition is influenced by both niche (deterministic) process, like selective filtering, and neutral (stochastic) processes, such as dispersal. The way these factors interact to determine microbial species composition remains unclear. For instance, high dispersal rates may overwhelm deterministic effects like selection, making community composition under the same conditions distinct. Alternatively, high dispersal rates could increase the opportunity for selection to occur, causing communities in similar conditions to become more similar. Empirical studies are often subject to sampling limitation, confounding factors, and methodological challenges, making it difficult to test the robustness of quantitative methods and identify general patterns. We aimed to test the relative influence of niche and neutral assembly processes and their interactions by examining how betadispersion, or the variance of replicate communities in a single treatment, is influenced by selection and dispersal. Using a trait-based model (DEMENT, Allison 2010), we varied dispersal rate (biomass and number of taxa immigrating each year), and the lignin:N of litter (representing a selection gradient). We exposed local communities derived from a constant regional pool of taxa to these conditions, and measured changes in betadispersion over time. 

Results/Conclusions

Overall, results showed significant effects of dispersal on community dynamics, and strong interactions between selection and the alteration of litter lignin:N. Dispersal was a stronger determinant of betadispersion, and this effect was even stronger under highly selective conditions (high lignin:N). Community composition was most divergent under a low-dispersal scenario, and at higher dispersal rates, communities became more similar. This suggests that dispersal limitation can drive differences in community composition, and can be more important than environmental selection for determining patterns of betadiversity over time and space. Simulations were also highly sensitive to many properties of dispersal that are unknown for microbial communities, like ratio between richness and biomass of the immigrant population, and emigration rates. Our study generates new hypotheses about the relative influence of deterministic and stochastic processes in microbial community assembly. In the future, these hypotheses could be tested by manipulating these factors in field studies.