Dissimilarity overlap curve analysis has shown that 'universality' is a common feature in many complex microbial communities, suggesting that the same taxa interact in a similar manner when shared between communities. We present evidence that arbuscular mycorrhizal fungi, common plant root symbionts, show universal community compositions in natural ecosystems and that this pattern is conserved even at larger spatial scales. However, universality was not detected in agricultural ecosystems potentially implying that agricultural symbiont communities are formed in a different manner.

Although ongoing research has revealed some of the main drivers behind global spatial patterns of microbial communities, spatio-temporal dynamics of these communities still remain largely unexplored. Here, we investigate spatio-temporal variability of both bacterial and eukaryotic soil microbial communities at local and intercontinental scales. We compare how temporal variation in community composition scales with spatial variation in community composition, and explore the extent to which bacteria, protists, fungi and metazoa have similar patterns of temporal community dynamics. All soil microbial groups displayed a strong correlation between spatial distance and community dissimilarity, which was related to the ratio of organism to sample size. Temporal changes were variable, ranging from equal to local between-sample variation, to as large as that between communities several thousand kilometers apart. Moreover, significant correlations were found between bacterial and protist communities, as well as between protist and fungal communities, indicating that these microbial groups change in tandem, potentially driven by interactions between them. We conclude that temporal variation can be considerable in soil microbial communities, and that future studies need to consider temporal variation in order to reliably capture all drivers of soil microbiome changes.

Creating accurate habitat suitability and distribution models (HSDMs) for soil microbiota is far more challenging than for aboveground organism groups. In this perspective paper, we propose a conceptual framework that addresses several of the critical issues holding back further applications. Most importantly, we tackle the mismatch between the broadscale, long-term averages of environmental variables traditionally used, and the environment as experienced by soil microbiota themselves. We suggest using nested sampling designs across environmental gradients and objectively integrating spatially hierarchic heterogeneity as covariates in HSDMs. Second, to incorporate the crucial role of taxa co-occurrence as driver of soil microbial distributions, we promote the use of joint species distribution models, a class of models that jointly analyze multiple species’ distributions, quantifying both species-specific environmental responses (i.e. the environmental niche) and covariance among species (i.e. biotic interactions). Our approach allows incorporating the environmental niche and its associated distribution across multiple spatial scales. The proposed framework facilitates the inclusion of the true relationships between soil organisms and their abiotic and biotic environments in distribution models, which is crucial to improve predictions of soil microbial redistributions as a result of global change.

No abstract

No abstract