Denitrification in restored and unrestored Danish streams

•Denitrification was measured in restored and unrestored stream sections.•Denitrification rates (DR) showed high spatial variability.•Restoration effects on DR may be minor compared to effects of local conditions.•Restoration may cause carbon limitation to denitrifiers.•Nitrate rich groundwater seepage can create denitrification hotspots.

Stream restoration often aims at mitigating nutrient pollution in aquatic ecosystems. However, despite recent research efforts, effects of restoration practices on in-stream nitrogen removal remain unclear. In this study, denitrification rates as well as factors controlling denitrification in unrestored and restored sections of two Danish streams (S1 and S2) were compared. The 15N isotope pairing technique was used to measure denitrification in situ. Denitrifier presence was analyzed by denaturing gradient gel electrophoresis (DGGE) and quantitative PCR of nitrite reductase (nirK and nirS) and nitrous oxide reductase (nosZ) genes. Denitrification rates were highly variable, with denitrification rates of 3106 μmol N m−2 h−1 in the unrestored section of S1, but no detectable denitrification in the restored section of S1, whereas in S2 restored and unrestored sections had similar denitrification rates of around 250 μmol N m−2 h−1. These large differences in denitrification rates were mainly due to differences in hydrologic conditions and sediment characteristics. High nitrate fluxes from upwelling groundwater created denitrification hotspots in the unrestored section of S1. Moreover, a lack of organic matter in the restored section of S1 likely caused a low abundance of denitrifiers and consequently no detectable denitrification. Our results indicate the importance of hydrology and sediment organic matter for stream nitrogen dynamics, which should be considered in restoration design.