The forces that structure ecological communities are the subject of much debate, particularly for microorganisms. Likewise, the relationships between microbial community assembly, biodiversity and ecosystem function are poorly understood. Here, I present a hypothesis describing a potential role for community assembly processes in microbial function, highlighting the potential importance of stochastic processes in decoupling the environment from processes of interest. I will then present data from empirical work, a metaanalysis and mathematical modeling to address three fundamental questions for assembly-biodiversity-function relationships: 1. What forces guide microbial community assembly? 2. Does microbial community structure matter for ecosystem function? 3. Do different assembly processes lead to different relationships between structure and function? Our work shows that both deterministic and stochastic processes are important for structuring soil bacterial communities during primary succession and after a major disturbance event. As well, our large metaanalysis reveals that data on functional components of microbial community composition can significantly improve models of ecosystem function based on environmental parameters alone. Finally, results from simulations using a combined community assembly-ecosystem function model demonstrate that stochastic processes can change the shape of biodiversity-ecosystem function relationships. Together, these data provide a testable framework for where, when and why microbial community assembly processes may matter for predictions of ecosystem function.

Diana Nemergut is an Associate Professor, INSTAAR / University of Colorado, Boulder; Department of Biology, Duke University