Management practices that reduce greenhouse gas (GHG) emissions while maintaining yields are an important option for reducing the global warming potential (GWP) of agriculture.  In flooded rice systems, nitrous oxide (N₂O) and methane (CH₄) are the primary GHGs emitted from soil.  Nitrogen (N) fertilizer is an important factor regulating N₂O emissions, thus proper N management is both critical for optimizing yields and reducing environmental impacts.  To identify N management practices that minimize GHG emissions per unit of grain yield produced (i.e. yield-scaled GWP), an on-farm experiment was conducted to quantify N₂O and CH₄ emissions during the rice growing season and winter fallow periods.  Fertilizer N in the form of aqua ammonia was applied at five N rates ranging from 0 to 260 kg N ha^-1.  GHG emissions were determined daily during specific management events (e.g. N application and field drainage periods), and every 7 d otherwise using the closed chamber technique.  Annual GHG emissions and yield-scaled GWP (N₂O + CH₄) were assessed with respect to N fertilizer rate, agronomic productivity, and N recovery efficiencies.  Results indicate that low N₂O emissions occurred regardless of N fertilizer rate when a permanent flood was maintained, with daily emissions remaining under 3.5 g N₂O-N ha^-1 day^-1.  During the growing season and fallow periods, CH₄ emissions were similar across N rates and ranged from 0 to approximately 3000 g CH₄-C ha^-1 day^-1, with the highest fluxes occurring midseason and during the drainage period before harvest.  Annual GWP was close to 3,000 kg CO₂ eq ha^-1 across treatments when N was applied.  As N rate increased yield-scaled GWP decreased from approximately 310 to 220 kg CO₂ eq Mg^-1, but significant differences were not observed between N rates.  These results suggest that optimal yields can be obtained at relatively constant yield-scaled GWP values when current N management recommendations are followed and N fertilizer inputs are closely matched with crop N demand.