The aims of this study were to use closed chambers to improve estimates of N2O-N losses from intensively managed grassland on poorly drained soils and to provide measurements for comprison with fluxes determined simultaneously using micrometeorological methods. A 10-ha field on clay soil in central Scotland received 185 kg NH4NO3-N ha-1 on April 3, 1992. Twenty-four closed chambers were installed, six in a 2--3-ha area grazed by cattle the previous summer, the remainder in an ungrazed area. Fluxes were measured regularly for 3 weeks. Nitrous oxide accumulation in the chambers was determined by gas chromatography. No flux was detected before fertilization. After fertilization, fluxes from the ungrazed and grazed areas were 153¿9 and 557¿107 g N2O-N ha-1 d-1, respectively (means and standard errors of all measurements). The individual fluxes ranged from 8 to 712 and 6 to 1519 g N2O-N ha-1 d-1, respectively, showing marked temporal and spatial variability and lognormal distributions. Fluxes peaked five days after fertilization and were one sixth of their maxima by April 24. Spatial differences observed initially were generally maintained.

Incubation of cores with 10% acetylene suggested that the N2O was produced by denitrification in the top 5 cm soil, and in situ soil N2O measurements confirmed that the concentration was highest close to the surface. In a regression model of the flux from the ungrazed area (including the pre- to postfertilization transition), air temperature, recent rainfall, and NO3--N could account for 52% of the temporal variability. The higher flux from the grazed area may have resulted from greater local heterogeneity of the surface soil in that area, arising from uneven compaction due to treading by livestock. The total N2O-N losses (1.7 and 5.1% of applied N from the ungrazed and grazed areas, respectively) confirm that fertilized grassland can contribute substantially to global N2O emissions. ¿ American Geophysical Union 1994