Defining Marine Microbial Biomes From Environmental And Dispersal Filtered Metapopulations

Energy and matter fluxes essential for all life1 are modulated by spatial and temporal shifts in microbial community structure resulting from environmental and dispersal filtering2,3, emphasizing the continued need to characterize microbial biogeography4,5. Yet, application of metapopulation theory, traditionally used in general ecology for understanding shifts in biogeographical patterns among macroorganisms, has not been tested extensively for defining marine microbial populations filtered by environmental conditions and dispersal limitation at global ocean scales. Here we show, from applying metapopulation theory on two major global ocean datasets6,7, that microbial populations exhibit core- and satellite distributions with cosmopolitan compared to geographically restricted distributions of populations. We found significant bimodal occupancy-frequency patterns (the different number of species occupying different number of patches) at varying spatial scales, where shifts from bimodal to unimodal patterns indicated environmental and dispersal filtering. Such bimodal occupancy-frequency patterns were validated in Longhursts classical biogeographical framework and in silico where observed bimodal patterns often aligned with specific biomes and provinces described by Longhurst and where found to be non-random in randomized datasets and mock communities. Taken together, our results show that application of metapopulation theory provides a framework for determining distinct microbial biomes maintained by environmental and dispersal filtering.