Soil microbial network complexity predicts ecosystem function along elevation gradients on the Tibetan Plateau

The contributions of soil microbial diversity for the multifunctionality of ecosystem functions remain little understood despite the well-known importance of soil microbial communities for regulating many ecosystem functions. Existing studies have typically focused on the role of univariate diversity or composition metrics, while ignoring that ecosystem processes are carried out via complex webs of co-occurrence networks among microbiome members. Here, the relationships of soil microbial diversity and network complexity (reflected by linkage density of inferred co-occurrence patterns) with multiple types of ecosystem functions (ecosystem multifunctionality) were assessed along a multi-elevation gradient (3755–5120 m) on the Tibetan Plateau. These functions included a suite of 18 variables related to soil nutrient status, soil carbon, nitrogen, and phosphorous cycles, plant growth, mitigation of greenhouse gases, and control of potential fungal plant pathogens in soils. The findings showed that the diversity, network complexity of both bacterial and fungal communities, as well as ecosystem multifunctionality significantly declined with increasing elevation. Although significant linear positive relationships were identified between diversity, network complexity of both microbial communities, and ecosystem multifunctionality, the relationships of network complexity had more positive relationships and explained more variance than diversity. These relationships remained robust when the confounding multifunctionality drivers (i.e., climatic, soil, and biotic factors) were simultaneously accounted for and when multiple multifunctionality approaches (i.e., multiple single functions, average multifunctionality, multithreshold, and multidimensional functions) were used. Further analysis showed that microbial diversity effects on multifunctionality were indirectly driven by microbial network complexity. This explicit empirical evidence shows that microbial network complexity enables a better prediction of ecosystem functions than the simplified microbial diversity metrics most studies applied in the past. Our results highlight the importance of co-occurrence networks as a key component of microbial biodiversity linking it with ecosystem multifunctionality.