Impact of redox and transport processes in a riparian wetland on stream water quality in the Fichtelgebirge region, southern Germany

Biologically mediated redox processes in the riparian zone, like denitrification, can have Substantially beneficial impacts on stream water quality. The extent of these effects, however. depends greatly oil the hydrological boundary conditions. The impact of hydrological processes oil a wetland's nitrogen sink capacity was investigated in a forested riparian fen which is drained by a first-order perennial stream. Here, we analysed the frequency distributions and time-series of pH and nitrogen, silica, organic carbon and oxygen concentrations in throughfall, soil solution, groundwater and stream water, and the groundwater levels and stream discharges from a 3-year period. During baseflow conditions, the stream was fed by discharging shallow, anoxic groundwater and by deep, oxic groundwater. Whereas the latter delivered considerable amounts of nitrogen (similar to 0-37 mg l(-1)) to the stream, the former was almost entirely depleted of nitrogen. During stormflow, near-surface runoff in the upper 30 cut soil layer bypassed the denitrifying zone and added significant amounts to the nitrogen load of the stream. Nitrate-nitro-en was close to 100% of deep groundwater and stream-water nitrogen concentration. Stream-water baseflow concentrations of nitrate, dissolved carbon and silica were about 1.6 mg l(-1), 4 mg l(-1) and 7.5 mg l(-1) respectively, and < 3 mg l(-1), < 10 mg l(-1) and > 4 mg l(-1) respectively during discharge peaks. In addition to that macroscale bypassing effect, there was evidence for a corresponding microscale effect: Shallow groundwater sampled by soil suction CLIPS indicated complete denitrification and lacked any seasonal signal Of Solute concentration, which was in contrast to piezometer samples from the same depth. Moreover, mean solute concentration in the piezometer samples resembled more that of suction-cup samples from shallower depth than that of the same depth. We conclude that the soil solution cups sampled to a large extent the immobile soil-water fraction. In contrast, the mobile fraction that was sampled by the piezometers exhibited substantially shorter residence time, thus being less exposed to denitrification, but predominating discharge of that layer to the stream. Consequently, assessing the nitrogen budget based on suction-cup data tended to overestimate the nitrogen consumption in the riparian wetland. These effects are likely to become more important with the increased frequency and intensity of rainstorms that are expected due to climate change. Copyright © 2006 John Wiley & Sons, Ltd.