Microbial Community Cohesion Mediates Community Turnover in Unperturbed Aquifers.

Microbial ecological processes are frequently studied in the presence of perturbations rather than in undisturbed environments, despite the relatively stable conditions dominating many microbial habitats. To examine processes influencing microbial community structuring in the absence of strong external perturbations, three unperturbed aquifers in Ohio (Greene, Licking, and Athens) were sampled over 2 years and analyzed using geochemical measurements, 16S rRNA gene sequencing, and ecological modeling. Redox conditions ranging from highly reducing to more oxidizing distinguished aquifer geochemistry across the three locations. Distinct microbial communities were present in each aquifer, and overall community structure was related to geochemistry, although community composition was more similar between the Athens and Licking locations. The ecological processes acting upon microbial assemblages within aquifers were varied; geochemical changes affected the Athens location, while time or some unknown factor affected Greene County. Stochastic processes, however, dominated the Licking aquifer, suggesting a decoupling between environmental fluctuations and community development. Although physicochemical differences might be expected to drive variable selection, dispersal limitation (inability to mix) explained differences between Athens and Licking. Finally, community complexity as measured by "cohesion" indicated that less-interconnected communities experienced higher turnover and were more likely to be affected by stochastic processes. Conversely, more-interconnected communities experienced lower turnover and susceptibility to homogenizing selection. Based upon these data, we support the hypothesis that unperturbed environments house dynamic microbial communities due to external and internal forces. IMPORTANCE Many microbial ecology studies have examined community structuring processes in dynamic or perturbed situations, while stable environments have been investigated to a lesser extent. Researchers have predicted that environmental communities never truly reach a steady state but rather exist in states of constant flux due to internal, rather than external, dynamics. The research presented here utilized a combined null model approach to examine the deterministic and stochastic processes responsible for observed community differences in unperturbed, groundwater ecosystems. Additionally, internal dynamics were investigated by relating a recently published measure of community complexity (cohesion) to ecological structuring processes. The data presented here suggest that communities that are more cohesive, and therefore more complex, are more likely affected by homogenizing selection, while less-complex communities are more susceptible to dispersal. By understanding the relationship between internal dynamics and community structuring processes, insight about microbial population development in natural systems can be obtained.