Constructed agricultural ponds and wetlands can reduce nitrogen loading from agriculture especially in areas where warm climate predominates. However, in cold climate temperature-dependency of microbiological processes have raised the question about the applicability of constructed wetlands in N removal. We measured in situ denitrification rates in a constructed agricultural pond using 15N-isotope pairing technique at ambient light and temperature throughout a year as well as diurnally. The field IPT measurements were combined with a wide set of potentially important explanatory data, including air temperature, photosynthetically active radiation, precipitation, discharge, nitrate plus other water quality variables, sediment temperature, oxygen concentration and penetration depth, diffusive oxygen uptake and sediment organic carbon. Denitrification varied, on average, diurnally between 12 and 314 μmol N m−2 h−1 and seasonally between 0 and 12409 μmol N m−2 d−1. Light and oxygen regulated the diel variation of denitrification, but seasonally denitrification was governed by a combination of temperature, oxygen and turbidity. The results indicated that the real N removal rate might be 30–35% higher than the measured daytime rates, suggesting that neglecting the diel variation of denitrification we may underestimate N removal capacity of shallow sediments. We conclude, that by following recommended wetland:catchment – size ratios, boreal agricultural ponds can efficiently remove nitrogen by denitrification in summer and in autumn, while in winter and in spring the contribution of denitrification might be negligible relative to the loading, especially with short residence time.