Environ. Sci. Technol. 2002, 36, 4547-4552
Soil Nitrogen Cycle Processes in Urban Riparian Zones PETER M. GROFFMAN* AND NATALIE J. BOULWARE Institute of Ecosystem Studies, Box AB, Millbrook, New York 12545 WAYNE C. ZIPPERER AND RICHARD V. POUYAT Northeastern Forest Experiment Station, U. S. Forest Service, Syracuse, New York 13210 LAWRENCE E. BAND Department of Geography, University of North Carolina, Chapel Hill, North Carolina 27514 MARK F. COLOSIMO U.S. Army Corps of Engineers, P.O. Box 1715, Baltimore, Maryland 21203
Riparian zones have been found to function as “sinks” for nitrate (NO3-), the most common groundwater pollutant in the U. S., in many areas. The vast majority of riparian research, however, has focused on agricultural watersheds. There has been little analysis of riparian zones in urban watersheds, despite the fact that urban areas are important sources of NO3- to nitrogen (N)-sensitive coastal waters in many locations. In this study, we measured stream incision, water table depths, and pools, production (mineralization, nitrification), and consumption (denitrification) of NO3in urban soils. Samples were taken from soil profiles (0100 cm) of three forested urban and suburban zones and one forested reference riparian zone in the Baltimore, Maryland metropolitan area. Our objectives were to determine (1) if stream incision associated with urbanization results in lower riparian water tables, and (2) if pools, production, and consumption of NO3- vary systematically with stream incision and riparian water table levels. Two of the three urban and suburban streams were more incised and all three had lower water tables in their riparian zones than the forested reference stream. Urban and suburban riparian zones had higher NO3- pools and nitrification rates than the forested reference riparian zone, which was likely due to more aerobic soil profiles, lower levels of available soil carbon, and greater N enrichment in the urban and suburban sites. At all sites, denitrification potential decreased markedly with depth in the soil profile. Lower water tables in the urban and suburban riparian zones thus inhibit interaction of groundwater-borne NO3- with near surface soils that have the highest denitrification potential. These results suggest that urban hydrologic factors can increase the production and reduce the consumption of NO3- in riparian zones, reducing their ability to function as sinks for NO3- in the landscape.
Introduction Riparian ecosystems have been shown to prevent the movement of pollutants from upland land uses to streams * Corresponding author phone: 845-677-7600; fax: 845-677-5976; e-mail:
[email protected]. 10.1021/es020649z CCC: $22.00 Published on Web 09/25/2002
2002 American Chemical Society
in many areas (1, 2). Most riparian research has focused on attenuation of nitrate (NO3-) in agricultural watersheds (3). Nitrate is the most common drinking water pollutant in U. S. groundwaters (4) and is an agent of eutrophication in coastal and marine waters (5). There has been relatively little work on riparian zones in urban watersheds, which is surprising given that these areas are frequently significant contributors of NO3- and other pollutants to receiving waters (6-8). Riparian NO3- attenuation functions depend on the capacity of riparian vegetation and microbial communities to intercept and process pollutants moving in surface runoff and groundwater flow (3). Denitrification, the anaerobic microbial conversion of NO3- to N gases (NO, N2O, and N2), is central to these functions. This process results in transfer of N to the atmosphere and is thus an important pathway for ecosystem losses of N (9). Regulation of denitrification in riparian zones is complex however, requiring interaction of groundwater-borne NO3- with anaerobic sites capable of supporting microbial activity (10, 11). There is still considerable uncertainty about which riparian zones are likely to support denitrification, and we are therefore unable to predict or manage riparian NO3- removal functions in many areas (12, 13). Several factors might cause riparian NO3- removal function to fail in urban watersheds. Hydrologic flow paths in these watersheds are greatly altered, with large amounts of water moving from uplands to streams as surface runoff, often in engineered structures (14, 15). Urban stream channels are often highly incised (16-18), which, in combination with reduced infiltration in impervious urban uplands, can reduce near-stream groundwater levels. Groundwater level is a key controller of the potential for interaction of groundwater-borne NO3- with organic-rich, microbially active near-surface soils (19, 20). Vegetation and soils in urban ecosystems are often highly disturbed, and few studies have examined soil microbial N-cycle processes in these ecosystems (21, 22). Low water tables and N enrichment in urban riparian zones could reduce NO3- consumption by denitrification and increase NO3- production by nitrification, an aerobic process that would be expected to be vigorous in relatively dry riparian areas subjected to high N loading from uplands or streams (23, 24). There is a strong need to evaluate riparian functions in urban watersheds, especially those adjacent to N-sensitive coastal waters (25, 5). In this study we evaluated soil N cycling processes in forested suburban and urban riparian zones in the Baltimore, Maryland metropolitan area. The work was part of the Baltimore Ecosystem Study, a new component of the U.S. National Science Foundation’s Long-Term Ecological Research (LTER) network (www.ecostudies.org/BES). We measured stream incision, water table depths, and pools, production, and consumption of NO3- in the soil profiles (0-100 cm) of three urban and suburban and one forested reference riparian zone. Our objectives were to determine (1) if stream incision associated with urbanization results in lower riparian water tables, and (2) if pools, production, and consumption of NO3- vary systematically with stream incision and riparian water table levels. Our goal was to understand these patterns and controls to aid in decision making about riparian assessment, management, and restoration to reduce NO3- delivery to the N-sensitive Chesapeake Bay, which receives runoff from the Baltimore metropolitan area. VOL. 36, NO. 21, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4547
TABLE 1. Land Use Setting, Watershed Area, Land Use Intensity, Stream Nitrate Concentrations, and Stream Incision (Depth) at Four Riparian Sites in Baltimore, MDa site
land use setting
watershed area (m2)
residential and use (%)
forest land (%)
stream nitrate (mg N L-1)b
stream incision (depth)
Pond Branch Glyndon Gwynnbrook Cahill
forested reference suburban (headwaters) suburban (downstream) urban
410 000 545 400 1 570 500 173 600
0 81 64 72
100 8 32 28
0.05 1.91 1.92 1.27
0.5 (0.1)b 0.5 (0.2)b 1.4 (0.2)ab 2.2 (0.7)a
a Two transects, consisting of two sampling locations 5 m from the stream bank on opposite sides of the stream, were located at each site. Values for stream incision (depth) are mean with standard error. Values followed by different letters within a column are significantly different at p < 0.05. b Volume-weighted concentration during water year 2000 (October 1999 - September 2000) at long-term weekly sampling stations near each riparian site. Data from Belt et al. (49).
TABLE 2. Soil Classification, Tree Basal Area, and Dominant Vegetation at Four Riparian Sites in Baltimore, MDa land use setting
site Pond Branch
forested reference
transect upper
soil classification fine-loamy, mixed, mesic Aquic Fragiudult
lower Glyndon
suburban (headwaters)
upper
coarse-loamy, micaceous, mesic Typic Dystrohrepts fine-loamy, mixed, mesic Typic Ochraquults
lower Gwynnbrook
suburban (downstream)
upper
0.89
Liriodendron tuilipifera (88%)
0.31
Nyssa sylvatica (4%) Liriodendron tuilipifera (46%) Acer rubrum (27%) Nyssa sylvatica (18%) Quercus alba (64%)
0.49
Prunus serotina (18%)
fine-loamy, mixed, mesic Aquic Fragiudult
1.56
1.67 0.68
Liriodendron tuilipifera (56%) Acer negundo (27%)
0.33 fluvent, frequently flooded
lower
Acer rubrum (13%) Fraxinus sp. (69%) Acer rubrum (17%) Ulmus rubra (10%) Acer rubrum (64%) Liriodendron tuilipifera (22%) Ailanthus altissima (9%) Acer rubrum (70%) Prunus pennsylvanica (28%) Platanus occidentalis (83%) Liriodendron tuilipifera (7%)
urban upper
a
dominant trees (% of basal area)
0.44
lower Cahill
tree basal area
Two transects, consisting of two sampling locations 5 m from the stream bank on opposite sides of the stream, were located at each site.
Experimental Section Sampling Sites. The BES LTER research project focuses on the Gwynns Falls watershed, a 17 150 ha catchment that traverses a gradient from the urban core of Baltimore, through older urban residential and suburban zones, rapidly suburbanizing areas, and a rural/suburban fringe (26). The project includes long-term stream monitoring; soil and vegetation analysis; spatial analysis of ecological, physical, and social variables; hydrologic, ecological, and social science modeling; and community education and outreach (www.ecostudies.org/BES). We located eight forested sites along four first- or secondorder steams in and around the Gwynns Falls watershed (Table 1). One site was in a completely forested catchment that served as a “reference” study area for the BES LTER. Two sites were in suburban areas of the watershed, one along the headwaters of the Gwynns Falls (less developed) and one along a tributary to the Gwynns Falls, which is farther downstream (more developed). The final, urban site was along a tributary to the Gwynns Falls in the urban core of the watershed. Residential areas near the suburban sites were approximately 50 years old, with small areas of much older (100-200 years) and much younger (1-5 years) development. Residential areas near the urban site were approximately 50-70 years old and more uniform than those in the suburban areas. All areas were served by sanitary sewers. 4548
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002
At each site, we marked two transects perpendicular to the stream, quantified stream channel depths, characterized vegetation, and established sampling plots 5 m from the stream edge on either side of the stream (Table 2). Stream channel depths were quantified by measuring the distance from the stream bank to the bottom of the stream channel, at approximately 2-m intervals along perpendicular transects running from points 5 m from the stream edge on either side of the stream. Overstory vegetation was characterized on a 400 m2 circular plot. Species and diameter at breast height (DBH) were recorded for all stems g2.5 cm DBH. Plot size was similar to that used by Brush et al. (27) to characterize the vegetation of Maryland. Water table wells (one per plot) constructed of 3.8-cm perforated PVC pipe were installed to 1.0 m below the ground surface. Wells were installed by hand augering to the desired depth, inserting the well casing, backfilling with clean sand to the top of the slots, then backfilling to within 15 cm of the surface with native fill, and filling the remaining hole with bentonite pellets. The wells were capped. Water table levels were measured at 1- to 4-week intervals from February 2000 to September 2001 using a tape measure attached to a simple electrical device that emits a signal when the water table is reached. Soil samples were taken at 0-5, 5-30, 30-50, 50-70, and 70-100 cm depth in July 2001 using a hand auger. Samples
were refrigerated and brought back to laboratories at the Institute of Ecosystem Studies in Millbrook, New York for analysis (described below). Analytical Methods. Samples were stored at 4 °C between sampling and analysis (less than 3 weeks). Soil samples were hand sorted and mixed, and held at field moisture for all analyses. Soil moisture content was determined by drying at 60 °C for 48 h (28). Soil organic matter content was determined by loss on ignition at 450 °C for 4 h (29). Amounts of inorganic N (NO3- and NH4+) in soil were determined by extraction with 2 M KCl followed by colorimetric analysis with a Perstorp Flow Solution Analyzer. Denitrification enzyme activity (DEA) (potential) was measured using the short-term anaerobic assay developed by Smith and Tiedje (30) as described by Groffman et al. (31). Sieved soils were amended with NO3-, dextrose, chloramphenicol, and acetylene, and were incubated under anaerobic conditions for 90 min. Gas samples were taken at 30 and 90 min, stored in evacuated glass tubes, and analyzed for N2O by electron capture gas chromatography. Rates of potential net N mineralization and nitrification, denitrification, and respiration were measured in a 10-day incubation of soils at room temperature. Soils were placed in 946-mL “mason” jars with lids fitted with septa for gas sampling. After 10 days, the headspace of the jars was sampled by syringe, and the gas samples were analyzed for carbon dioxide (CO2) by thermal conductivity gas chromatography. The jars were then opened, flushed, and re-sealed, and acetylene (10 kPa) was then added to the headspace of the jars. After 24 h, gas samples were taken again and analyzed for nitrous oxide (N2O) by electron capture gas chromatography. Mineralization was calculated as the accumulation of total inorganic N, nitrification was calculated as the accumulation of NO3-, and respiration was calculated as the accumulation of CO2 over the course of the 10-day incubation. Denitrification was calculated as the accumulation of N2O over the course of the 24-hour incubation with acetylene. This approach to measuring denitrification avoids problems with NO3- pool depletion that can occur when acetylene, which inhibits nitrification, is used, because NO3- pools are allowed to accumulate for 10 days before denitrification is measured (31). However, the method likely underestimates denitrification because the disturbed soils in a jar are likely more aerobic than intact soils in the field. Statistical Analysis. Sites were compared using one-way analysis of variance with Fisher’s least significant difference test to compare the four site means. Person product-moment correlations were used to explore relationships between variables. The SAS statistical program was used for all analyses (32).
Results The urban stream had a significantly (p < 0.05) deeper channel than the forested reference and suburban (headwaters) streams (Table 1). Depth of the suburban (downstream) stream was not significantly different from those of the other streams. There were marked differences in water table depth and dynamics between the forested reference and the urban and suburban sites (Figure 1). Water tables were relatively close to the surface and constant in the forested reference site and much deeper and more variable in the urban and suburban sites. Over all dates, the water table at the forested reference site was significantly (p < 0.05) closer to the soil surface than at the other sites. Despite the differences in stream channel depth and water table, there were few differences in denitrification potential (DEA) among the sites (Table 3). Subsurface soils had much lower DEA than surface soils, but there was measurable denitrification potential at all depths (Table 3). There were
FIGURE 1. Water table depth at four riparian sites in Baltimore, MD from February 2000 to September 2001. Each panel compares one suburban or urban site with the same forested reference site. Values are mean of four wells along two transects at each site. significant correlations between DEA and soil NO3- levels (r ) 0.70, p < 0.0001), potential net nitrification rate (r ) 0.61, p < 0.0001), potential net mineralization rate (r ) 0.52, p < 0.0004), and respiration (r ) 0.31, p < 0.04). The forested reference site had much lower pools of NO3and potential net nitrification rates than the urban and suburban sites (Figure 2, differences were significant (p < 0.05) in the top three depths). Although the forested reference site had low levels of NO3- and nitrification throughout the soil profile, all of the urban and suburban sites had significant VOL. 36, NO. 21, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4549
TABLE 3. Denitrification Enzyme Activity in Soil Profiles at Four Riparian Sites in Baltimore, MDa forested reference
depth (cm)
suburban suburban (headwaters) (downstream) kg-1
0-5 5-30 30-50 50-70 70-100
456 (152)b 42 (5)b 25 (10)a 20 (3)a 23 (3)a
µg N 1664 (164)a 95 (50)b 60 (40)a 18 (1)a 27 (1)a
urban
h-1
586 (124)b 29 (7)b 34 (1)a 50 (24)a 193 (137)a
2196 (63)a 421 (149)a 105 (52)a 50 (6)a 77 (25)a
a Values are mean (standard error) of two riparian transects at each site (n ) 2). Each transect consisted of two sampling locations 5 m from the stream bank on opposite sides of the stream. Values followed by different letters within a row are significantly different at p < 0.05.
pools and production of NO3-. Rates of NO3- production in the urban and suburban sites were high enough (e.g., 0.5 mg N kg-1 d-1 in the suburban headwaters site) to account for the observed NO3- pools (e.g., 3.0 mg N kg-1 in the suburban headwaters site). Potential net nitrification rates were correlated with NO3- pools (r ) 0.90, p < 0.0001) and potential net mineralization rates (r ) 0.81, p < 0.0001). Denitrification rates were very low relative to NO3- pools and nitrification rates (µg rather than mg N kg-1 d-1) and did not vary with depth or site. Denitrification rates were correlated with NH4+ pools (r ) 0.37, p < 0.01) and soil moisture content (r ) 0.34, p < 0.02). The forested reference site had much higher (p < 0.05) levels of soil organic matter and respiration at the surface depth than the urban and suburban sites (Figure 3). The forested reference site had a well-developed organic surface horizon (forest floor), but the other sites did not. Subsurface organic matter and respiration did not differ among the sites. Respiration and organic matter were highly correlated (r ) 0.94, p < 0.0001). Respiration was also correlated with NH4+ pools (r ) 0.44, p < 0.002) and soil moisture content (r ) 0.74, p < 0.0001).
Discussion Changes in Soil N Cycle Processes. We hypothesize that down-cutting and incision of the stream channel by storm flows, and reduced infiltration in the uplands associated with urbanization, has lowered water tables in these riparian zones. A lower water table creates a more aerobic soil profile, increasing rates of NO3- production by nitrification, an aerobic process (23, 24), and reducing rates of NO3consumption by denitrification. In addition to affecting soil N cycling processes, lower water tables alter groundwater flow paths from uplands towards the stream, causing groundwater-borne NO3- to move through subsurface material with relatively low denitrification potential (see below). These changes have the net effect of increasing the production and reducing the consumption of NO3- in urban riparian zones, decreasing their potential to function as NO3- sinks in urban watersheds. In addition to increases in aerobic conditions, two other factors may have contributed to the high NO3- pools and production that we observed in the urban and suburban riparian zones: organic matter depletion and N enrichment. The forested reference site had high levels of organic matter and respiration in surface soils. Organic matter increases NO3- consumption in soils by fueling heterotrophic denitrifiers and stimulating uptake of inorganic N by heterotrophic microbes, i.e., immobilization (24). Even though organic matter and respiration were higher only in the surface soils of the forested reference site, the surface horizon may supply C to microbes lower in the soil profile via leaching of dissolved organic C. The lower levels of organic matter in the urban 4550
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002
FIGURE 2. Soil nitrate (top), potential net nitrification (middle), and denitrification (bottom) in soil profiles at four riparian sites in Baltimore, MD. Values are mean (standard error) of two riparian transects at each site. Each transect consisted of two sampling locations 5 m from the stream bank on opposite sides of the stream. and suburban soils were not unexpected. An increase in aerobic conditions due to lowered water tables would be expected to lead to a depletion of organic matter in the soil profile (24). Rates of soil N cycle processes in the urban and suburban riparian zones were also likely increased by N enrichment from groundwater and/or surface water. Although we have no data on groundwater NO3- concentrations, streamwater in the urban and suburban watersheds was highly enriched in NO3-. Riparian soil processes are known to respond to N enrichment (23, 33). Enrichment does not necessarily reduce the sink potential of riparian zones because increases in denitrification can balance increases in nitrification (23). However, in urban and suburban riparian zones with low water tables, denitrification is inherently inhibited and the potential for loss of the riparian sink is increased.
FIGURE 3. Soil organic matter (top) and respiration (bottom) in soil profiles at four riparian sites in Baltimore, MD. Values are mean of two riparian transects at each site. Each transect consisted of two sampling locations 5 m from the stream bank on opposite sides of the stream. It is interesting to note that denitrification potential in surface and subsurface soils was not reduced in the urban and suburban sites relative to that of the forested reference site, despite the changes in stream depth and water table level associated with urbanization. This lack of difference was probably caused by two factors: the N enrichment discussed above, and the fact that the urban and suburban sites were relatively “intact” forest sites, without extensive soil disturbance or invasion by exotic plant species. These sites had relatively high levels of surface soil organic matter, which strongly influences denitrification potential as discussed above. Urban Hydrology Effects on Riparian Zones. There is a long history of research on how urbanization can result in stream channel changes. Conversion of agricultural and forest lands to residential, commercial, and industrial uses (i.e., urbanization), produces an increase in both volume and peak flow of runoff (34, 35). The increased flows can increase channel erosion, and ultimately, increase sediment loading to downstream reaches (36). The increases in runoff and changes in sediment supply to a stream result in stream channel disequilibrium. Over time, the stream channel adjusts to the changed flow regime, and erosion and sediment supply are reduced (16, 18, 37). A variety of physical changes are associated with a streams’ response to these changes in runoff. Increased entrainment and transport of sediment from the bed and banks results in changes in sediment erosion and deposition patterns, including both incision and sedi-
ment island or bar formation, which promote channel instability and lead to channel widening (36). For example, Nanson and Young (34) found stream cross-section area increased by a factor of 2 or 3 relative to pre-development conditions. Channel enlargement occurred through bank incision and widening. Booth (17) demonstrated similar responses to urbanization in Pacific Northwest streams. Although the effects of urbanization on stream channels have been relatively well studied, there has been no analysis of the effects of these changes on riparian water table levels and biological functions, even though there has been extensive research on the importance of shallow subsurface groundwater flow to riparian groundwater NO3- removal function. If NO3- is moving from uplands toward streams in shallow groundwater (e.g., < 50 cm from the soil surface), there is high potential for removal of this NO3- by plant roots and microbes that are abundant in near surface soils (19, 38, 39). If NO3- is moving in deep groundwater, or in surface runoff, the potential for removal is greatly reduced. Our data are consistent with the idea that urban and suburban streams have deeper channels and lower water table levels in the riparian zone than streams in completely forested watersheds. In the suburban and urban sites, the water table was relatively deep in the soil profile where denitrification potential was 1 or 2 orders of magnitude lower than that at the surface. The low denitrification potential of subsurface soil compared to that of surface soil is consistent with the findings of many previous studies (20, 39-42). Subsurface denitrification is frequently limited by a lack of carbon to support denitrifiers, which are primarily heterotrophic (43, 44). In our urban riparian zones, groundwater, and any NO3- that it is carrying, is clearly moving through material with low denitrification potential, which likely greatly reduces the NO3- removal function of these riparian zones. Our data suggest that, in the most extreme case, urban riparian zones can become “hydrologically isolated” from uplands and streams in urban watersheds. If water movement from uplands to streams is primarily as stormwater flow (mostly in pipes) or as deep groundwater flow, riparian zones have less opportunity to influence the quality of this water and have a reduced ability to function as NO3- sinks in urban watersheds. Urban streams are highly dynamic in time and space, and further research is needed to develop predictive relationships between urban stream dynamics, near-stream water table levels, and riparian denitrification. Urban stream incision is strongly influenced by watershed sediment production, which is controlled by land disturbance associated with construction and agricultural activities (16). Highly incised streams are most likely to be found in older urban areas with stable land cover. In addition to sediment and stormwater loads, stream channel depth is influenced by other factors that can vary widely within and between urban areas. Stream age, sediment texture, and human disturbance can all affect stream channel and water table depths (45). For example, our suburban (headwaters) site had a relatively shallow stream channel, but deep water table depths. This site was extensively disturbed during installation of sanitary sewer lines (which run adjacent to the stream) several decades ago, and the main channel was split into two shallow smaller channels. The fact that water table levels were low at this site suggests that other factors, e.g., reduced infiltration in uplands, influence near-stream water table levels in urban watersheds. Infiltration of water into buried storm and sanitary sewer lines can also contribute to low water tables in urban riparian zones, with these lines functioning like the artificial drainage that is used to lower water tables to facilitate agricultural production in soils with poor natural drainage. In arid regions, irrigation and wastewater discharges might create an opposite VOL. 36, NO. 21, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4551
“urban effect”, acting to increase near-stream water table levels. The use of septic systems, which discharge into groundwater, could also lead to increases in groundwater levels, especially in watersheds where human water supply comes from outside the watershed. Evaluating riparian denitrification function in specific urban watersheds clearly requires consideration of multiple hydrologic, soil, and geomorphic factors to determine whether there is a hydrologic connection between uplands, the riparian zone, and the stream. Implications for Restoration of Urban Watersheds and Streams. The relatively high denitrification potential of the surface soils that we observed in the urban and suburban sites suggests that there is potential for restoring or creating riparian NO3- removal functions in urban watersheds by altering hydrology. If we can reconnect riparian zones to uplands and streams, either by restoring shallow subsurface flow or by creating wetlands to detain and process stormwater, we may be able to enhance, restore, or create an important NO3- sink in urban watersheds. In agricultural watersheds, restoring or creating “sinks” has been considered to be a more promising approach to reducing NO3- delivery to N-sensitive coastal waters than source control in several cases (46-48). Altering hydrology in highly urbanized watersheds is a great challenge, but as has been found in agricultural watersheds, it may be easier to engineer and restore sinks than to control highly diffuse sources. In urban watersheds, these sources, which range from home lawns to leaky sewers and septic systems to illegal discharges, may be particularly hard to define and control.
(14) (15) (16) (17) (18) (19) (20) (21) (22) (23) (24) (25) (26)
(27) (28)
(29)
Acknowledgments Alex Kalejs was a central contributor to this research, carrying out field sampling, laboratory analysis, and data handling. We also thank Jessica Hopkins, Alan Lorefice, and Emilie Stander for help with these tasks. This paper is a contribution to the Baltimore Ecosystem Study and was funded by National Science Foundation grant DEB 97-14835. Funding support also came from the USDA Forest Service Research Work Unit (NE-4952), Syracuse, NY.
Literature Cited (1) Gilliam, J. W. J. Environ. Qual. 1994, 25, 896-900. (2) Correll, D. L. Buffer zones and water quality protection: General principles; Haycock, N. E., Burt, T. P., Goulding, K. W. T., Pinay, G., Eds.; Quest Environmental: Harpenden, UK, 1997; pp 7-20. (3) Hill, A. R. J. Environ. Qual. 1996, 25, 743-755. (4) U.S. Environmental Protection Agency. National Pesticide Survey: Nitrate. Office of Water, Office of Pesticides and Toxic Substances: Washington, DC, 1990. (5) Howarth, R. W.; Billen, G.; Swaney, D.; Townsend, A.; Jaworski, N.; Lajtha, K.; Downing, J. A.; Elmgren, R.; Caraco, N.; Jordan, T.; Berendse, F.; Freney, J.; Kudeyarov, V.; Murdoch, P.; Zhu, Z. L. Biogeochem. 1996, 35, 75-139. (6) Makepeace, D. K.; Smith, D. W.; Stanley, S. J. Crit. Rev. Environ. Sci. Technol. 1995, 25, 95-139. (7) Miller, C. V.; Denis, J. M.; Ator, S. W.; Barkebill, J. W. J. Am. Water Res. Assoc. 1997, 33, 1155-1171. (8) Grimm, N. B.; Grove, J. M.; Pickett, S. T. A.; Redman, C. L. BioScience 2000, 50, 571-584. (9) Firestone, M. K. Biological denitrification. In Nitrogen in Agricultural Soils; Stevenson, F. J., Ed.; American Society of Agronomy: Madison, WI, 1982; pp 289-326. (10) Cirmo, C. P.; McDonnell, J. J. J. Hydrol. 1997, 199, 88-120. (11) Jacinthe, P. A.; Groffman, P. M.; Gold, A. J.; Mosier, A. J. Environ. Qual. 1998, 27, 156-164. (12) Lowrance, R.; Altier, L. S.; Newbold, J. D.; Schnabel, R. R.; Groffman, P. M.; Denver, J. M.; Correll, D. L.; Gilliam, J. W.; Robinson, J. L.; Brinsfield, R. B.; Staver, K. W.; Lucas, W.; Todd, A. H. Environ. Manag. 1997, 21687-21712. (13) Correll, D. L. The current status of our knowledge of riparian buffer water quality functions. In Proceedings of the International Conference on Riparian Ecology and Management in Multi-Land 4552
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 21, 2002
(30) (31)
(32) (33) (34) (35) (36) (37) (38) (39) (40) (41) (42) (43) (44) (45) (46) (47) (48) (49)
Use Watersheds; Wigington, P. J., Jr., Beschta, R. L., Eds.; American Water Resources Association: Portland, Oregon, 2000; pp 5-10. Field, R. Water Resources Bulletin, American Water Resource Association 1985, 2, 21. Scheuler, T. Water. Protect. Technol. 1994, 1, 100-111. Wolman, M. G. Phys. Geog. 1967, 49, 385-395. Booth, D. B. Water Resour. Bull. (now J. Am. Water Resour. Assoc.) 1990, 26, 407-417. Henshaw, P. C.; Booth, D. B. J. Am. Water Resour. Assoc. 2000, 36, 1219-1236. Nelson, W. M.; Gold, A. J.; Groffman, P. M. J. Environ. Qual. 1995, 24, 691-699. Groffman, P. M.; Howard, G.; Gold, A. J.; Nelson, W. M. J. Environ. Qual. 1996, 25, 1309-1316. McDonnell, M. J.; Pickett, S. T. A. Ecology 1990, 71, 1231-1237. DeKimpe, C. R.; Morel, J.-L. Soil Sci. 2000, 165, 31-40. Hanson, G. C.; Groffman, P. M.; Gold, A. J. Ecol. Applic. 1994, 4, 750-756. Paul, E. A., Clark, F. E. Soil Microbiology and Biochemistry; Academic Press: New York, 1996. Valiela, I.; Collins, G.; Kremer, J.; Lajtha, K.; Geist, M.; Seely, B.; Brawley, J.; Sham, C. H. Ecol. Applic. 1997, 7, 358-380. Doheny, E. Index of Hydrologic Characteristics and Data Resources for the Gwynns Falls Watershed, Baltimore County and Baltimore City, Maryland; USGS Open-File Report; United States Geological Survey: Baltimore MD, 1999; pp 99-213. Brush, G. S.; Lenk, C.; Smith, J. Ecol. Monogr. 1980, 50, 77-92. McInnes, K. J.; Weaver, R. W.; Savage, M. J. Soil water potential. In Methods of Soil Analysis, Part 2 - Microbiological and Biochemical Properties; Weaver, R. W., Ed.; Soil Science Society of America: Madison WI, 1994; pp 53-58. Nelson, D. W.; Sommers, L. E. Total carbon, organic carbon, and organic matter. In Methods of Soil Analysis, Part 3 Chemical Methods; Sparks, D. L., Ed.; Soil Science Society of America: Madison WI, 1996; pp 961-1010. Smith, M. S.; Tiedje, J. M. Soil Biol. Biochem. 1979, 11, 262-267. Groffman, P. M.; Holland, E.; Myrold, D. D.; Robertson, G. P.; Zou, X. Denitrification. In Standard Soil Methods for Long-Term Ecological Research; Robertson, G. P., Bledsoe, C. S., Coleman, D. C., Sollins, P., Eds.; Oxford University Press: New York, 1999; pp 272-288. SAS. SAS/STAT User’s Guide, Release 6.03. SAS Institute Incorporated: Cary, NC, 1988. Groffman, P. M. Curr. Top. Wet. Biogeochem. 1994, 1, 15-35. Nanson, G.; Young, R. J. Hydrology 1981, 52, 239-251. Arnold, C.; Boison, P.; Putton P. J. Geol. 1982, 90, 155-166. Wolman, M. G.; Schick, A. P. Water Resour. Res. 1967, 3, 451464. Roberts, W. P.; Pierce, J. P. Sediment. Geol. 1974, 12, 179-197. Simmons, R. C.; Gold, A. J.; Groffman, P. M. J. Environ. Qual. 1992, 21, 656-665. Hill, A. R.; DeVito, K. J.; Campagnolo, S.; Sanmugadas, K. Biogeochemistry 2000, 51, 193-223. Hiscock, K. M.; Lloyd, J. W.; Lerner, D. N. Water Res. 1991, 9, 1099-1111. Korom, S. F. Water Resour. Res. 1992, 28, 1657-1668. Groffman, P. M.; Gold, A. J.; Simmons, R. C. J. Environ. Qual. 1992, 21, 666-671. Bradley, P. M.; Fernandez, M., Jr.; Chapelle, F. H. Environ. Sci. Technol. 1992, 12, 2377-2381. Starr, R. C.; Gillham, R. W. Ground Water 1993, 31, 934-947. Pizzuto, J. E.; Hession, W. C.; McBride, M. Geology 2000, 28, 79-82. Gren, I. M. Ecol. Engin. 1995, 4, 153-162. Gren, I. M.; Soderqvist, T.; Wulff, F. J. Environ. Manag. 1997, 51, 123-143. Mitsch, W. J.; Day, J. W.; Gilliam, J. W.; Groffman, P. M.; Hey, D. L.; Randall, G. W.; Wang, N. BioScience 2001, 51, 373-388. Belt, K. T.; Groffman, P. M.; Band, L. E.; Law, N. L.; Fisher, G. T. Ecology, submitted for publication.
Received for review March 18, 2002. Revised manuscript received August 12, 2002. Accepted August 13, 2002. ES020649Z
