Process-based modelling of nitrous oxide emissions from different nitrogensources in mown grassland

The process-based Pasture Simulation Model (PaSim 2.5) has been extended to simulate N2O production and emission from grassland caused by nitrogen inp uts from different sources. The model was used to assess the influence of m anagement on N2O emissions, such as the effect of shifts in the amount and timing of fertilizer application. Model performance has been tested against season-long field measurements at two different field sites. Simulation re sults agreed favourably with measured N2O emission and soil air concentrati ons, except during an extremely wet period at one site when grass growth wa s very poor. The results of short-term and long-term simulation runs demons trated the potential of the model to estimate N2O emission factors under va rious conditions. During the first growing season, simulated emissions from organic fertilizers were lower than from synthetic fertilizers because mor e of the nitrogen was used to build up soil organic matter. The relative di fference between the fertilizer types became larger with increasing applica tion rate. The difference between fertilizer types was smaller at steady-st ate when higher soil organic matter content from repeated application of or ganic fertilizer over time led to enhanced mineralization and N2O emissions . The dependence of simulated N2O emissions on N input was close to linear at low, but non-linear at high fertilization rates. Emission factors calcul ated from the linear part of the curve suggested that the factors used in t he current IPCC method underestimate the long-term effects of changes in fe rtilizer management. Furthermore the simulations show that N2O emissions ca used by nitrogen inputs from the decomposition of harvest losses and from b iological fixation in grassland can be considerable and should not be negle cted in national emission inventories.