Forested riparian zones may protect waterways from receiving excess N from adjacent agricultural lands by immobilizing N in plants or microbes, or through denitrification. If denitrified, excess N is emitted either as harmful nitrous oxide or benign N2 gas. Therefore the size, structure and activity of the denitrifying community may play important roles in determining the net effect of riparian zones on the overall environmental impact of excess agricultural N. We assessed the size and structure of prokaryotic denitrifying communities, measured soil physiochemical characteristics and estimated denitrification along a gradient of grazed dairy pasture to forested riparian zone buffering the Manawatu River in New Zealand. We assessed denitrifier community structure by T-RFLP of the nirS and nosZ genes. We assessed denitrifier community size by quantitative PCR (qPCR) of these genes and total bacterial numbers by qPCR of the ribosomal polymerase (rpoB) gene. Strong gradients in microbial biomass, gene abundances and potential denitrification were primarily driven by differences in soil texture between pasture and riparian zones, with smaller scale patterns in soil properties emerging among slope positions with each zone. We found that nirS and nosZ communities responded strongly, though sometimes independently, to these gradients. nosZ abundance and community structure were significantly correlated to large scale patterns edaphic characters that distinguished pasture and riparian zone soils, but were not significantly correlated to denitrifier activity. By contrast, nirS abundance responded in a unimodal fashion to soil texture, leading to strong patterns of abundance and community structure associated with the different slope positions. Furthermore, nirS community structure was significantly correlated to denitrification rates and end-products. Finally, we found that the evenness of denitrifying communities was uncorrelated denitrification parameters and was negatively correlated to the proportional abundance of denitrifiers. Taken together, our data suggest that small numbers of genotypes become dominant and drive higher rates of denitrification in favourable ‘hotspots’ in the toe-slopes of pastures. Conversely, the low overall community evenness and low proportional abundance of complete denitrifiers in the riparian zone soils suggested a limited capacity for riparian buffers to remove excess N through denitrification and a propensity for incomplete denitrification to lead to N2O emissions. These results contribute to our understanding of the role of riparian zones determining the fate of excess agricultural N and highlight the importance of careful N management especially in pastures that occur on coarse textured soils.