Terrestrial nitrogen cycle simulation with a dynamic global vegetation model

A global scale Dynamic Nitrogen scheme (DyN) has been developed and incorporated into the Lund-Posdam-Jena (LPJ) dynamic global vegetation model (DGVM). The DyN is a comprehensive process-based model of the cycling of N through and within terrestrial ecosystems, with fully interactive coupling to vegetation and C dynamics. The model represents the uptake, allocation and turnover of N in plants, and soil N transformations including mineralization, N₂ fixation, nitrification and denitrification, NH₃ volatilization, N leaching, and N₂, N₂O and NO production and emission. Modelled global patterns of site-scale nitrogen fluxes and reservoirs are highly correlated to observations reported from different biomes. The simulation of site-scale net primary production and soil carbon content was improved relative to the original LPJ, which lacked an interactive N cycle, especially in the temporal and boreal regions. Annual N uptake by global natural vegetation was simulated as 1.084 Pg N yr⁻¹, with lowest values <1 g N m⁻² yr⁻¹ (polar desert) and highest values in the range 24-36.5 g N m⁻² yr⁻¹ (tropical forests). Simulated global patterns of annual N uptake are consistent with previous model results by Melillo et al. The model estimates global total nitrogen storage potentials in vegetation (5.3 Pg N), litter (4.6 Pg N) and soil (>=67 Pg as organic N and 0.94 Pg as inorganic N). Simulated global patterns of soil N storage are consistent with the analysis by Post et al. although total simulated N storage is less. Deserts were simulated to store 460 Tg N (up to 0.262 kg N m⁻²) as NO₃⁻, contributing 80% of the global total NO₃⁻ inventory of 580 Tg N. This model result is in agreement with the findings of a large NO₃⁻ pool beneath deserts. Globally, inorganic soil N is a small reservoir, comprising only 1.6% of the global soil N content to 1.5 m soil depth, but the ratio has a very high spatial variability and in hot desert regions, inorganic NO₃⁻ is estimated to be the dominant form of stored N in the soil.