Patchiness in Groundwater Nitrate Removal in a Riparian Forest

Our ability to identify and manage riparian sites for groundwater nitrate (NOAh) removal is limited by uncertainty surrounding the relative importance of plant uptake vs. microbially mediated removal processes. Microcosm studies often demonstrate negligible transformation rates in the subsoil of riparian forests, even in situations where groundwater well networks showed substantial groundwater NO~removal during the winter and a decline in dissolved oxygen (DO) in ambient groundwater moving through the site. We hypothesize that microcosm studies may miss groundwater transformations that occur within microsites, that is, "hotspots" of riparian subsoils. We created mesocosms of large (15 cm diam. × 40 cm length), undisturbed cores from the seasonally saturated zone of poorly drained (PD) and moderately well drained (MWD) sandy soils from a forested riparian area in southern New England. We dosed the mesocosms for 130 d with ambient groundwater amended with NO~-N and Br-. Changes in the NO~--N/Brratios were used to calculate groundwater NO~-N removal rates. The PD treatment demonstrated substantial groundwater NOj--N removal rates. The PD mesocosms contained patches of dark-stained material that often surrounded roots in various stages of decay. The dry mass of patches in the PD treatment ranged from 0.07 to 1.4% of the mesocosms. The MWD treatment contained no patches and exhibited no groundwater NO~-N removal. Further investigations on the relationships between the extent of subsurface patchiness, water table dynamics and plant characteristics might yield fruitful insights into the management of vegetated riparian zones for groundwater NO~--N removal. C controversy surrounds the fate of nitrate-nitrogen (NO~--N) in groundwater. Several studies suggest that NO;is relatively conservative in groundwater (Keeney, 1986; Bradley et al., 1992; Starr and Gillham, 1993), but several studies indicate that NO; may undergo marked transformations in portions of aquifers. In particular, groundwater NO; removal has been noted in the shallow groundwater below vegetated riparian zones (Lowrance et al., 1984; Jacobs and Gilliam, 1985; Lowrance, 1992; Nelson et al., 1995; Hill, 1996; Verchot et al., 1997). Riparian zones are defined functionally as transition areas that have the capacity to affect the exchange of energy and matter between uplands and surface waters (Gregory et al., 1991). Because groundwater contributes substantially to the base flow of many lower order streams, the fate of groundwater NO~ in riparian zones can have a major effect on the export of NO; from watersheds. Our ability to identify and manage riparian sites for groundwater NO;removal is limited by uncertainty surrounding the relative importance of plant uptake vs. microbially mediated removal processes (Hill, 1996). Riparian settings with permeable surface soils, shalA.J. Gold, W.R. Wright, and R.H. Puffer, Dep. of Natural Resources Science, University of Rhode Island, Kingston, R102881; P.A. Jacinthe and P.M. Groffman, Institute of Ecosystem Studies, Box AB, Millbrook, NY 12545. Received 17 Apr. 1997. *Corresponding author ([email protected]). Published in J. Environ. Qual. 27:146-155 (1998). low groundwater, and extended groundwater retention times (i.e., days vs. hours) are the most likely locations for groundwater NO;removal (Hill, 1996). But these settings have been found to exhibit substantial intrasite variation in shallow groundwater NO; removal rates (Cooper, 1990; Simmons et al., 1992; Schipper et al., 1993; Nelson et al., 1995). Starr and Gillham (1993) suggested that denitrification rates in groundwater may be inversely related to the depth of the water table. Soil characteristics, particularly the depth to redoximorphic features can be a useful indicator of water depth and groundwater NO; removal within riparian zones. At the upland edge, a riparian zone may contain welldrained (WD) and moderately well-drained (MWD) soil drainage classes with water table depths >1 m from the surface. Near the stream, the water table depths of PD and very poorly drained (VPD) soils may be within 0.4 m from the surface. We have found that PD riparian soils consistently demonstrate more groundwater NO~removal capacity than MWD riparian soils (Simmons et al., 1992; Nelson et al., 1995). Much of our knowledge of groundwater NO3removal capacity of riparian locations has been obtained through field studies based on groundwater monitoring well networks (Hill, 1996). In situ studies are cumbersome for isolating the magnitude and dynamics of individual removal processes such as plant uptake, microbial immobilization, and denitrification. In addition to the need for process-level research, there is a role for techniques that can provide a more rapid assessment of the variation in removal capacities that results from different hydrogeomorphic settings, plant communities, groundwater depth, and seasonal conditions. Our research group (Groffman et al., 1992, 1996) and others (Lowrance, 1992) have compared the shallow groundwater NO3removal estimated from monitoring well networks with microcosm studies of microbial NO; transformations, that is, denitrification and/or immobilization, in soil derived from the same depths and locations. In all cases the microcosms demonstrated negligible transformation rates, even in situations where groundwater well network studies showed substantial groundwater NO;removal during the winter and a decline of DO in ambient groundwater moving through the site. The microcosm results suggest that either plant uptake is the predominant process involved in NO; removal or that microcosm results may not capture a sufficient sample volume to characterize microbial processes in the groundwater of riparian zones (Groffman et al., 1996; Smith et al., 1996). Microcosm studies often rely on <50 g of sample to represent field conditions. Even with intensive sampling, this sample size may miss "hot spots" of microbial activity (Parkin et al., 1987; Christensen et al., 1990a). Abbreviations: PD, poorly drained; MWD, moderately well drained; VPD, very poorly drained; DO, dissolved oxygen.