Main content area

Constraining N cycling in the ecosystem model LandscapeDNDC with the stable isotope model SIMONE

Denk, Tobias R. A., Kraus, David, Kiese, Ralf, Butterbach‐Bahl, Klaus, Wolf, Benjamin
Ecology 2019 v.100 no.5 pp. e02675
ammonium, biogeochemical cycles, ecosystems, environmental factors, grasslands, greenhouse gas emissions, microorganisms, model validation, models, nitrates, nitrification, nitrogen, nitrous oxide, pollution, pollution control, soil, soil organic nitrogen, stable isotopes, water content, Switzerland
The isotopic composition (ic) of soil nitrogen (N) and, more recently, the intramolecular distribution of ¹⁵N in the N₂O molecule (site preference, SP) are powerful instruments to identify dominant N turnover processes, and to attribute N₂O emissions to their source processes. Despite the process information contained in the ic of N species and the associated potential for model validation, the implementation of isotopes in ecosystem models has lagged behind. To foster the validation of ecosystem models based on the ic of N species, we developed the stable isotope model for nutrient cycles (SIMONE). SIMONE uses fluxes between ecosystem N pools (soil organic N, mineral N, plants, microbes) calculated by biogeochemical models, and literature isotope effects for these processes to calculate the ic of N species. Here, we present the concept of SIMONE, apply it to simulations of the biogeochemical model LandscapeDNDC, and assess the capability of ¹⁵N‐N₂O and, to our knowledge for the first time, SP, to constrain simulated N fluxes by LandscapeDNDC. LandscapeDNDC successfully simulated N₂O emission, soil nitrate, and ammonium, as well as soil environmental conditions of an intensively managed grassland site in Switzerland. Accordingly, the dynamics of ¹⁵N‐N₂O and SP of soil N₂O fluxes as simulated by SIMONE agreed well with measurements, though ¹⁵N‐N₂O was on average underestimated and SP overestimated (root‐mean‐square error [RMSE] of 8.4‰ and 7.3‰, respectively). Although ¹⁵N‐N₂O could not constrain the N cycling process descriptions of LandscapeDNDC, the overestimation of SP indicated an overestimation of simulated nitrification rates by 10–59% at low water content, suggesting the revision of the corresponding model parameterization. Our findings show that N isotope modeling in combination with only recently available high‐ frequency measurements of the N₂O ic are promising tools to identify and address weaknesses in N cycling of ecosystem models. This will finally contribute to augmenting the development of model‐based strategies for mitigating N pollution.