Climate prediction relies on land surface carbon cycle simulations in Earth System Models (ESMs). Key uncertainties in this regard are how ecosystems respond to elevated atmospheric CO2 and how the relevant mechanisms are implemented in ESMs. In this study, we survey quantitative and qualitative evidences of how elevated CO2affects plant nitrogen (N) uptake and soil N supply and compare these evidences against eleven state-of-the-art ecosystem models at two Free Air Concentration Enrichment (FACE) sites (Oak Ridge and Duke).
Results/Conclusions:
Eleven models that participated in the model intercomparison project for these sites predicted that, under elevated CO2, the intra-system N cycle was accelerated (higher bulk soil net N mineralization) and became more conservative (less NO3- leaching loss, higher Nitrogen Use Efficiency (NUE)). In contrast, observations indicate that bulk soil N mineralization, leaching, and NUE were not affected by CO2 enrichment. Instead, changing root resource foraging strategies, accelerated rhizosphere N cycling, and mycorrhizal fungi processes supported higher plant N demand. Therefore, the models that correctly predicted the magnitude of plant N uptake and soil N supply at these two FACE sites did so for the wrong reasons. Using existing models to project CO2 fertilization effects on terrestrial ecosystems carbon and nitrogen dynamics and carbon-climate feedbacks are therefore uncertain and likely unreliable. To address these issues, we propose a new ESM land modeling framework that represents the observed mechanisms affecting carbon and nutrient interactions in the plant-soil interface. This framework is designed to account for tradeoffs between the large-scale dynamics required for global predictions and the fine-scale processes observed in the field.