Nitrous oxide (N 2O) emissions are a large proportion of the agriculture sector's contribution to the greenhouse gas inventory of most developed countries. The spatial and temporal variability of N 2O emissions from agricultural soils has long been considered the main factor limiting our ability to estimate N 2O emissions, particularly the emissions associated with the spring snowmelt period. Tower and aircraft-based flux measurement systems and a process-based model were used to quantify N 2O emissions for four years (2000, 2001, 2003 and 2004) in an agricultural area of eastern Canada, near Ottawa, where a corn-soybean crop rotation dominates. A tower-based system, which relies on the flux gradient technique, provided diurnal N 2O emissions at a field scale. An aircraft-based system, which relies on the relaxed eddy accumulation technique, provided N 2O emissions for two similar agricultural regions and the DeNitrification and DeComposition (DNDC) model was used to estimate daily N 2O emissions at a regional scale. In most cases, aircraft-based N 2O emissions measurements were comparable for the two agricultural regions. Corresponding tower-based measurements which were collected over a field in the Ottawa area showed similar emission patterns to the aircraft-based measurements but in some cases the tower-based emissions were larger, as expected. This is because the footprint of aircraft-based measurements always incorporated a significant amount of crops such as soybean and other types of vegetation which do not receive additional nitrogen fertilization as well as waterlogged areas that do not emit N 2O. While in three of the four years, the tower-based measurements were made over a tile drained field where nitrogen fertilizer had been applied the previous year. The N 2O emissions patterns after planting were also similar for both aircraft and tower-based systems, but again they were slightly larger for the tower-based system. Aircraft-based N 2O flux measurements are also compared to the N 2O emissions obtained using the most recent version of the process-based model DNDC. Tests showed that DNDC gave comparable N 2O emissions estimates for the measurement period as a whole, but was not always able to correctly predict the timing of peak emissions.