TESTING HIGH-RESOLUTION NITROUS-OXIDE EMISSION ESTIMATES AGAINST OBSERVATIONS USING AN ATMOSPHERIC TRANSPORT MODEL

Global monthly estimates of N2O emissions were used to prescribe a thr ee-dimensional atmospheric transport model. The simulated N2O surface concentrations in the northern hemisphere (NH) are in general agreemen t with atmospheric observations, about 1 ppb higher than in the southe rn hemisphere (SH). The results suggest that N2O concentrations over s trong source regions in continental interiors are much higher than tho se over oceans. Simulated N2O concentrations in the MI, vary strongly with the season, while the measured concentrations oscillate strongly without seasonality. The modelled results for the SH show much less se asonal variation, which is more consistent with atmospheric measuremen t data. Most atmospheric N2O monitoring stations are not well situated for verifying the simulated seasonality in atmospheric N2O. There is only one continental site. In addition, atmospheric N2O observations f rom different studies are sometimes contradictory. The difference betw een model results and observations may be the result of local scale pu lses of N2O acting on a much shorter timescale than the time step of 1 month of the emission inventory. Such N2O pulses may cause oscillatio ns of the N2O concentration at the monitoring sites which dominate the large-scale seasonality. It is also possible that source and sink pro cesses other than those included in this study are operating. The tran sport model does not include a description of the downward stratospher e-troposphere flux of N2O which is greater in the NH than in the SH. T he emission inventory does not account for soil N2O consumption in N-I imited temperate ecosystems, and N2O emissions during autumn, thaw per iods, and from snow covered soils. These potential errors and possible underestimation of N2O emissions from combustion sources may exaggera te the simulated seasonal trends.