Hydrological variability, organic matter supply and denitrification in the Garonne River ecosystem

1. Groundwater nitrate contamination has become a worldwide problem as increasing amounts of nitrogen fertilisers are used in agriculture. Alluvial groundwater is uniquely juxtaposed between soils and streams. Hydrological connections among these subsystems regulate nutrient cycling. 2. We measured denitrification using an in situ acetylene‐block assay in a nitrate‐contaminated portion of the Garonne River catchment along a gradient of surface water–ground water mixing during high (snowmelt) and low flow. 3. During high flow (mid‐April to early June) the water table rose an average of 35 cm and river water penetrated the subsurface to a great extent in monitoring wells. Denitrification rates averaged 5.40 μgN₂O L^?1 min^?1 during the high flow period, nearly double the average rate (2.91 μgN₂O L^?1 min^?1) measured during base flow. This was driven by a strong increase in denitrification in groundwater under native riparian vegetation. Nitrate concentrations were significantly lower during high flow compared with base flow. Riparian patches had higher dissolved organic carbon concentrations that were more aromatic compared with the gravel bar patch closest to the river. 4. Multiple linear regression showed that the rate of denitrification was best predicted by the concentration of low molecular weight organic acids. These molecules are probably derived from decomposition of soil organic matter and are an important energy source for anaerobic respiratory processes like denitrification. The second best predictor was per cent surface water, reflecting higher denitrification rates during spring when hydrological connection between surface water and ground water was greatest. 5. Our results indicate that, while denitrification rates in Garonne River alluvium were spatially and temporally variable, denitrification was a significant NO₃ sink during transport from the NO₃‐contaminated floodplain to the river. DOC availability and river–floodplain connectivity were important factors influencing observed spatial and temporal patterns.