Conversion of landscapes to large-scale agriculture has substantially increased the loading of bioavailable nitrogen (N) to stream networks through extensive artificial drainage and fertilizer application. Floodplain restoration may enhance N cycling in agricultural stream systems by increasing residence time of floodwaters in contact with bioreactive surfaces that retain or remove excess N. Microbially-mediated denitrification potentially plays a significant role in constructed floodplains by permanently removing excess N through conversion of bioavailable nitrate (NO3-N) to dinitrogen gas, either as N2O or N-2. Restoring channelized streams via the construction of inset floodplains can increase the total denitrification capacity of agricultural watersheds, but little is known about the abiotic factors that control the proportion of NO3-N that is converted to the potent greenhouse gas N2O versus N-2 (i.e., N2O yield). We used an in-situ static core design to assess the importance of constructed floodplain age, inundation time, and carbon (C) availability on total denitrification rates and N2O yield. Novel use of membrane-inlet mass spectrometry (MIMS) allowed us to simultaneously measure N2O and N-2 to capture total denitrification and the proportion of bioavailable N converted to each end-product. Floodplain age did not influence total denitrification rates, but rather denitrification increased with the duration of floodplain inundation until C limitation occurred at approximately 24 h. In addition, we found that N2O yields from floodplain soils were higher than those reported for other aquatic systems. Finally, while floodplain restoration in agricultural streams generally increases N retention at the watershed scale, regardless of the restoration age, the impact of added floodplains on N2O emissions should be considered.