Predicting microbial nutrient limitations from a stoichiometry-based threshold framework

While ecological stoichiometry-based theories and methodologies have been successfully applied to identify nutrient limitations in plankton and plants (e.g., leaf nutrient resorption ratios), determining nutrient limitations in soil microorganisms by them has great uncertainties. Establishing stoichiometric thresholds that define specific nutrient limitations for soil microbial communities has been a major challenge. Here, we present a threshold framework that predicts the limitations of carbon (C), nitrogen (N), and phosphorus (P) for microbial communities by leveraging the convergence of elemental stoichiometry from soils to microorganisms. Combined with a widely used model of microbial nutrient limitations (i.e., ecoenzyme vector model), the theoretical thresholds derived from this framework are validated by empirical thresholds identified through analysis of a comprehensive global dataset (n = 3277). Notably, we find that the commonly assumed threshold of 45�� in the ecoenzyme vector model overestimates microbial P limitation and underestimates microbial N limitation. By contrast, a vector length of 0.61 (unitless) and an angle of 55�� in the vector model are more reliable thresholds identifying microbial C and N/P limitations, respectively. This framework, with a fundamental understanding of microbially-driven stoichiometric convergence in plant-soil-microbe systems, offers a practical and general avenue for identifying microbial nutrient limitations in soils.