Restorations of former cranberry farms (“bogs”) aim to re-establish native wetland vegetation, promote cold water habitat, and attenuate nitrogen (N) delivery to coastal waters. It is unclear, though, how elements of restoration design such as microtopography, groundwater interception, and plant communities affect N removal via denitrification. In a recently restored riparian cranberry bog with created microtopography, we compared denitrification potential, nitrous oxide (N2O) yield of denitrification (ratio of N2O:N2O + N2 gases), in situ N2O fluxes, soil chemistry, and plant communities at the highest and lowest elevations within 20 plots and at four side-channel groundwater seeps. Denitrification potential was > 2 × greater at low elevations, which had plant communities distinct from high elevations, and was positively correlated with plant species richness (Spearman’s rho = 0.43). Despite detecting high N2O yield (0.86 ± 0.16) from low elevation soils, we observed small N2O emissions in situ, suggesting minimal incomplete denitrification even in saturated depressions. Groundwater seeps had an order of magnitude higher denitrification potentials and 100–300 × greater soil NO3− concentrations than the typically saturated low elevation soils. Groundwater seeps also had high N2O yield (1.05 ± 0.15) and higher, but spatially variable, in situ N2O emissions. Our results indicate that N removal is concentrated where soils interact with NO3–rich groundwater, but other factors such as low soil carbon (C) also limit denitrification. Designing restoration features to increase groundwater residence time, particularly in low lying, species rich areas, may promote more N attenuation in restored cranberry bogs and other herbaceous riparian wetlands.