Gang Fu, Wei Sun
The stability of soil microbiomes is critical for ecosystem functioning under climate change, yet its assessment is confounded by the overlooked problem of observational temporal scale dependency. Here, we introduce a multi-temporal window framework to resolve this problem. Applied to a decade-long warming experiment, our approach reveals that the perceived stability of bacterial and fungal communities nonlinearly decays with observational temporal scale, and short windows systematically overestimate it. Crucially, we document a temporal scale-driven mechanistic shift. Community stability shifts from species resistance to compensatory asynchrony once the window exceeds a threshold. This transition occurs over a broader temporal scale range for fungi than for bacteria. Our work establishes that microbiome stability is an intrinsically temporal scale-dependent property and provides a scalable, bioinformatic-friendly framework that challenges conventional single-temporal-scale assessments. This paradigm is critical for accurately predicting the fate of soil carbon and other microbiome-governed functions in a warming world. Soil microbiome community stability is a dynamic and temporal scale-dependent emergent property and is maintained by a temporal scale-driven shift from reliance on species resistance to compensatory asynchrony, according to a decade-long warming experiment in Damxung County, North Tibetan Plateau.