Nutrient limitation is a key source of uncertainty in predicting terrestrial carbon (C) uptake. Models have begun to address this uncertainty by including nitrogen (N) dynamics; however, phosphorus (P) which can also limit or co-limit net primary production (NPP) in many ecosystems, is absent or poorly parameterized in most models. To meet this challenge, we integrated P dynamics into a fully coupled, cutting-edge plant N model (Fixation & Uptake of Nitrogen: FUN) that predicts the C cost of N uptake from soil based on the cost of allocating C to ectomycorrhizal (ECM) or arbuscular mycorrhizal (AM) roots. We incorporated the direct C cost of P uptake, as well as an additional C and N cost of synthesizing phosphatase enzymes to extract P from soil, into a new model formulation (herein, FUN-P). We confronted and validated FUN-P against empirical estimates of canopy, root, and soil P pools in temperate forests. We then ran a model experiment to examine the extent to which the costs of P acquisition varied as a function of mycorrhizal association, and N and P availability in soils.
Results/Conclusions
Across 45 temperate forest plots that vary in the distribution of AM and ECM associated trees in southern Indiana, we found that FUN-P accurately predicted empirical estimates of P translocation to leaves. Moreover, the addition of P costs improved the ability of FUN-P (relative to the baseline model) to capture observed patterns of C allocation to root exudation and mycorrhizal biomass. In the model experiment, there was a clear interaction between mycorrhizal association and nutrient content. Regardless of N availability, AM trees took up more P than ECM trees at both high and low levels of soil P owing to both the greater P demand as well as the lower C cost of P uptake by AM fungi due to their high affinity P transporters. For ECM trees, there was an interactive effect of soil N and P levels on the C cost of P uptake. ECM trees had higher C costs of P uptake under low soil N than under low soil P, reflecting the larger indirect N costs of using enzymes to liberate P from soil organic matter. Collectively, the integration of P into FUN-P provides a novel framework for modeling how interactions between the C-N-P cycles belowground impact the ability of plants to acquire nutrients to support NPP.