Traditionally, agricultural nitrous oxide (N₂O) emission of New Zealand has been measured using chambers or lysimeters, and micrometeorological flux measurement experiments have been very few. Since micrometerological flux measurement systems have the advantage of measuring spatially integrated flux values for longer time periods compared to measurements made using chambers, development and verification of such a system was needed for New Zealand's agro-meteorological conditions. In this study, efficacy of such a combined flux gradient (FG) - eddy covariance (EC) micrometeorological flux measurement system is verified by continuously measuring N₂O fluxes from some control and mitigated agricultural plots of New Zealand. The control fields had natural N₂O emission, whereas, the mitigated plots were treated with chemicals to reduce N₂O emission. In this combined FG-EC method, the turbulent eddy diffusivities were estimated using the Monin-Obukhov (M-O) similarity theory based parameterization (where diffusion velocity `dhp' was used) and a thermal approach (where eddy diffusivity `kht' was used) from the EC measurements. These transfer coefficients (kht and dhp) along with the measured N₂O concentration differences were then fitted to the traditional FG equation to compute final flux values. As the primary objective of this study, measured fluxes from two different seasons and from two approaches were compared for consistency and then verified against published results. Under this wider objective of verification of the FG-EC micrometeorological method of N₂O flux estimation, this research thesis addresses three key issues: (i) assessment of error propagation in the measured flux through the eddy diffusivity - to understand the random error dynamics of the system and to estimate precision of the overall method, (ii) quantification and separation of N₂O source area emission rates from adjacent plots - to identify the contribution of an individual plot to the measured flux when multi-plot fluxes were measured from sources with different biogenic characters, and (iii) quantification of the effect of animal grazing and mitigation on the measured flux and actual emission rate of N₂O - to assess robustness of the FG-EC micrometeorological system. As a fourth objective of this study, (iv) new scaling properties of a turbulence surface layer model of a convective atmosphere is investigated as an alternative to the standard M-O similarity theory, as significant questioning of the M-O theory has been reported in some recent publications. Results from the verification experiment showed that the daily measured flux values obtained from this combined micrometeorlogical system for control plots varied between 0-191.9 and 0-491.8 gN₂O-N.ha⁻¹.day⁻¹ for autumn and spring experiments, respectively, for the parameterization method. Similarly, the daily mean flux values were found to be 10.9 ± 0.98 and 11.7 ± 0.57 gN₂O-N.ha⁻¹.day⁻¹ for the autumn and spring seasons, respectively. All these values were found to be of the same order of previously reported values in the literature and found to verifying that this FG-EC system works well under a range of meteorological conditions within a defined error range. Therefore, when the propagated random error was computed in the final flux value using kht and dhp, the mean relative error in kht was found to be higher than the mean relative error in dhp, irrespective of stability. From a Monte-Carlo type simulation of the random error, it was found that the maximum error can be up to 80% for kht irrespective of stability, and 49% and 35% for dhp respectively for stable (1/L ≥ 0, where L is Obukhov length) and unstable (1/L < 0) atmosphere. Errors in the concentration differences were estimated based on the minimum resolvable estimates from the gas analyzer and the associated random errors were found to be 6% and 8% for unstable and stable conditions. Finally, the total mean random error in the N2Oflux values was found to be approximately of the order of 9% and 12% for the parameterization method for unstable and stable conditions, respectively, and 16.5% for the thermal method, irrespective of stability. Objective (ii) of this research was addressed by developing a `footprint fraction' based inverse footprint method. Results of the footprint analysis method were assessed, first, by comparing footprint fractions obtained from both an analytical footprint model and a `forward' simulation of a backward Lagrangian stochastic (bLs) model; and second, by comparing the source area emission rates of a control plot obtained from the footprint analysis method and from the `backward' simulation of the bLs model. It was observed that the analytical footprint fractions were realistic as they compared well with the values obtain from the bLs model. The actual emission rates were found to be on average 2.1% higher than the measured flux values for the control plots. On average 4.3% of the measured fluxes were found to be contributed by source areas outside of the field domain. Again, the proposed footprint method of emission rate estimation was found to work well under a wider range of atmospheric stability, as the inverse footprint model and bLs model based emission rates were found to correlate well (0.70 and 0.61 for autumn and spring, respectively) with a 99% statistical significance. Similarly when the effect of grazing on the N₂O fluxes was considered, a 90% enhancement in the flux values was observed after grazing, followed by a decreasing trend in fluxes. However, contrary to existing knowledge of mitigation of N₂O flux by an inhibitor, this study found no statistically significant effect of mitigation in the pastoral emission of N₂O. Error accumulation, lesser soil N₂O production potential and/or inefficiency of the FG-EC method was conjectured to be reason/s for such discrepancy and some alternative convective boundary layer turbulence scaling was tested. Separate field measurement data, including the vertical profile measurements of the convective boundary layer and sonic anemometer measurements within the surface layer were used for this purpose. The spectral analysis of the vertical wind component, temperature and heat flux revealed that this new model of the convective boundary layer, which explains atmospheric boundary layer turbulence in terms of some nonlocal parameters, is more suitable than the traditional Monin-Obukhov similarity theory based model of atmospheric turbulence where the atmospheric flow properties are local. Therefore, it can be concluded that this new model of turbulence might provide the framework for a newer model of flux estimation in future. Overall, the FG-EC model of N₂O flux estimation method seems to work well within a certain error range. However, more field applications of this FG-EC method are needed for different agro-meteorological conditions of New Zealand before this method is accepted as a standard method of flux estimation, particularly, inefficiency in detecting the effect of mitigation should be tested. Development of an alternative flux gradient model which includes nonlocal atmospheric surface parameters might also be considered as a future research objective.