Coupling the resource stoichiometry and microbial biomass turnover to predict nutrient mineralization and immobilization in soil

Abstract The mineralization of organic nitrogen (N) and phosphorus (P) by the soil microbial biomass, as well as microbial immobilization of their mineral forms, can be predicted from differences between the stoichiometry of organic substrate and the nutrient demand of the microbial biomass. The accuracy of such predictions, however, decreases when the nutrient demand of microbial biomass changes in response to nutrient limitation or excess. We quantified net N and P mineralization/immobilization along gradients of organic substrate stoichiometry in a short-term (2-day) incubation experiment. Gradient of organic substrate stoichiometry (water extractable organic N and P concentrations) was created by mixing soils from two spruce forest soils (from both litter and organic topsoil horizons) at five different ratios. Biological predictors (i.e., microbial carbon (C) use efficiency and microbial biomass C, N, and P) of net nutrient mineralization/immobilization were quantified, and theoretical N and P mineralization/immobilization rates were predicted using known stoichiometric relationship. Measured net N and P immobilization was lower than that predicted. Extended mathematical modelling in combination with stable isotope analysis showed that the capability of microbial community to reduce its demand for external nutrients was responsible for the difference between the predictions and observations. Active part of microbial community instantly recycled N from decaying part of microbial community and very likely utilized internal P sources (i.e. polyphosphates) when the abundance of N and P in available organic compounds was insufficient. Our results suggest that the N recycling from dead microbial biomass and the internal microbial P sources warrant further investigation. Including these mechanisms in soil biogeochemical models based on ecological stoichiometry principles could improve their predictive accuracy.