Background/Question/Methods   Recently, resource managers have combined restoration strategies with wetland creation as a best management practice (BMP) in an effort to reduce the affects of increased nutrient loading in urban landscapes. Although, riparian wetlands have been shown to be particularly effective “sinks” for nitrogen, the vast majority of riparian nitrogen removal research has been in agricultural and forested watersheds. There has been relatively little work on riparian wetland function in urban watersheds. We investigated the question, “What is the variation and magnitude of denitrification in constructed urban wetlands compared to undeveloped forested wetlands. Three wetland types were evaluated in the Baltimore Metropolitan area: urban oxbow and constructed stormwater wetlands, and undeveloped forested wetlands. The oxbow wetlands formed in relict stream channels abandoned during geomorphic stream restoration and the stormwater wetlands were created below outfalls collecting runoff. Denitrification rates were measured via the 15N-enriched nitrate “push–pull” method in wetlands sediments during the summer and winter of 2008. We evaluated differences in denitrification across study sites, wetland type, distance, and season (all interactions included) using GLM for an unbalanced ANOVA followed by Tukey's post hoc test for differences among means (accepted significant at the alpha = 0.05 level).

Results/Conclusions   The in-situ denitrification push-pull test detected spatial and temporal variability in denitrification with rates across all wetland types ranging from < 0.1 to 193.2 μg N•kg soil-1•d-1. Sediment denitrification rates in urban wetlands were similar to forested wetlands during the summer and winter, but were generally higher during the winter. The highest denitrification rates were observed in constructed storm water wetlands (147 ± 28.9 μg N•kg soil-1•d-1) compared to urban oxbow wetlands (100 ± 10.9 μg N•kg soil-1•d-1) and forested wetlands (106 ± 32.3 μg N•kg soil-1•d-1) (mean ± standard error). Water column NO3- concentrations in the wetlands ranged from undetectable to 11 mg/L, and were lower than concentrations in adjacent streams, suggesting that the wetlands are “sinks” for NO3-. Denitrification rates were high relative to NO3- standing stock in the water column, suggesting that denitrification underlies the NO3- sink function of the wetlands. Our results suggest that urban wetlands have the potential to function as NO3- sinks in urban landscapes. The high denitrification rates observed in summer and winter suggest these wetlands are sinks for NO3- year round. This study also provides information for resource managers on how to design stream restoration projects to optimize denitrification in urban areas.