Spatial and Temporal Variation in Groundwater Nitrate Removal in a Riparian Forest

We quantified nitrate (NO~-) removal rates from groundwater in a red maple (Acer rubrum L.) riparian forest subjected to NOjdosing. The site was in Southern New England on soils classified as sandy mixed mesic Haplaquept soils and contained somewhat poorly (SPD) and poorly drained (PD) soils. The specific objectives were to examine groundwater NO£ removal rates within a riparian forest with respect to: (i) soil drainage class; (il) depth below water tabl e; and (ili time of year. We created 16 experimental dosing/monitoring stations at two depths along three soil drainage class transects (SPD, SPD/ PD, PD). We added solution containing NO~and Brcontinuously for II mo to a dosing well at each station. Groundwater was monitored at sampling wells 0.6 m downgradient of the dosing well. Nitrate removal rates were determined by coupling changes in the NO~-/Brratio with groundwater flux estimates from each experimental station. Although located just 20 m downgradient, the PD transects had substantially higher NO£ removal rates and lower dissolved oxygen than the SPD transects. Both the SPD/PD and PD transects had considerable NO~" -N removal capacity throughout the upper 1.5 m oftbe groundwater (16-46 lag kg-~ d-l). Rates were not significantly influenced by temperature. The scale of variation in removal rates suggests that high resolution soil and groundwater maps may be needed when riparian forests are to be used for water quality management. ]~ vom~sa~s have been shown to have a high capacity to prevent the movement of nitrate (NO~-) in surface runoff and groundwater flow from upland agricultural and residential land uses into streams (Lowrance et al., 1984; Peterjohn and Correll, 1984; Jacobs and Gilliam, 1984). Nitrate, a federally regulated drinking water pollutant and a prime agent of eutrophication in marine and freshwater ecosystems, is the most commonly detected groundwater pollutant in the USA (Ryther and Dunstan, 1971; Keeney, 1987; USEPA, 1990). Maintenance and restoration of riparian zones to control NO~pollution is a key component of several state and federal nonpoint-source pollution control efforts (Vellidis et al., 1993; NRC, 1993). Despite widespread claims about the role of riparian forests in nonpoint-source pollution control, several aspects of their function related to spatial and temporal variability are unclear. Although several studies have demonstrated that riparian forests located on wet organic soils have a high ability to attenuate groundwater NO~(Lowrance et al., 1984; Peterjohn and Correll, 1984; Jacobs and Gilliam, 1985; Pinay and Decamps, 1988; Simmons et al., 1992; Jordan et al., 1993; Haycock and Pinay, 1993), riparian forests on more well-drained mineral soils appear to be much less effective (Lowrance, 1992; Simmons et al., 1992). In riparian forests dominated by wetlands, a greater proportion of groundwaterborne NO~moves within the biologically active zone of W.M. Nelson, Wehran EMCOM NE, Wallingford, CT 06492; A.J. Gold, Dep. of Nat. Resources Sci., Univ. of Rhode Island, Kingston, RI 02881; and P.M. Groffman, Institute of Ecosystems Studies, Box AB, Millbrook NY 12545. Received 27 June 1994. *Corresponding author (agold@ edcserv.edc.uri, edu). Published in J. Environ. Qual. 24:691-699 (1995). the soil and is thus susceptible to uptake by plants and microbes. However, the site-specific characteristics that influence attenuation processes in riparian forests dominated by wetland/upland transition zones are much less clear (Simmons et al., 1992; Haycock and Pinay, 1993). There is much uncertainty about the mechanisms of groundwater NO~removal (Lowrance, 1992; Groffman et al., 1992). Several studies have observed NO~disappearance from groundwater during winter months (Lowrance et al., 1984; Simmons et al., 1992; Haycock and Pinay, 1993; Haycock and Burt, 1993; Jordan et al., 1993). In these situations, denitrification and/or microbial immobilization in the subsurface must be responsible for observed NO~removal. While subsurface denitrification has been observed in several studies (Trudell et al., 1986; Slater and Capone, 1987; Smith and Duff, 1988; Francis et al., 1989; Obenhuber and Lowrance, 1991), other studies have found the potential for denitrification in the subsurface to be low or nonexistent, usually due to a lack of available carbon (C) (Parkin and Meisinger, 1989; Groffman et al., 1992; Bradley et al., 1992; McCarty and Bremner, 1992; Johnson and Wood, 1992; Yeomans et al., 1992; Starr and Gillham, 1993). In a previous study in glacial outwash soils in Rhode Island, we measured low rates of in vitro subsurface denitrification in sites that demonstrated a high capacity for in situ groundwater NO~removal in winter (Simmons et al., 1992; Groffman et al., 1992). Lowrance (1992) reported a similar conundrum in the Southeastern Coastal Plain in Georgia. Our previous work also found high spatial variation in NO~removal within riparian forests, with removal ranging from <20% in some relatively dry upland/wetland transition zone soils to >90% in very poorly drained or hydric soils. In this study, we quantified in situ groundwater NO~removal rates in a natural, undisturbed, riparian forest receiving nutrient enrichment. The specific objectives were to examine groundwater NO3removal rates within a riparian forest with respect to: (i) soil drainage class; (ii) depth below groundwater table; and (iii) time of year. In a companion study completed at the same location, laboratory microcosms were used to characterize and quantify microbial processes potentially involved in NO~removal at the site (G. Howard, unpublished data). Our experimental system was designed to allow us to calculate actual NO~removal rates rather than simply the percentage of NO~removal. Calculation of removal rates facilitates applications to watershed scale analysis, as well as the comparison of in situ hydrologic studies with laboratory-based microcosm microbial studies.