JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 7 May 2007
Published in J Environ Qual 36:801-814 (2007)
DOI: 10.2134/jeq2006.0270
© 2007 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Young, E. O.
Right arrow Articles by Briggs, R. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Young, E. O.
Right arrow Articles by Briggs, R. D.
Agricola
Right arrow Articles by Young, E. O.
Right arrow Articles by Briggs, R. D.
Related Collections
Right arrow Field-Scale Studies
Right arrow Nutrient Cycling
Right arrow Ground Water Quality

TECHNICAL REPORTS

Landscape and Watershed Processes

Nitrogen Dynamics among Cropland and Riparian Buffers

Soil-Landscape Influences

Eric O. Younga,* and Russell D. Briggsb

a Dep. of Plant and Soil Science, Univ. of Vermont, Hills Agricultural Building, 105 Carrigan Dr., Burlington, VT 05405
b Dep. of Forest and Natural Resources Management, SUNY College of Environmental Science and Forestry, 1 Forestry Dr., Syracuse, NY 13210

* Corresponding author (eoyoung{at}uvm.edu)

Received for publication July 11, 2006.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Nitrate (NO3–) leaching to ground water poses water quality concerns in some settings. Riparian buffers have been advocated to reduce excess ground water NO3– concentrations. We characterized inorganic N in soil solution and shallow ground water for 16 paired cropland-riparian plots from 2003 to 2005. The sites were located at two private dairy farms in Central New York on silt and gravelly silt loam soils (Aeric Endoaqualfs, Fluvaquentic Endoaquepts, Fluvaquentic Eutrudepts, Glossaquic Hapludalfs, and Glossic Hapludalfs). It was hypothesized that cropland N inputs and soil-landscape variability would jointly affect NO3– leaching and transformations in ground water. Results showed that well and moderately well drained fields had consistently higher ground water NO3– compared to more imperfectly drained fields receiving comparable N inputs. Average 50-cm depth soil solution NO3– and ground water dissolved oxygen (DO) explained 64% of average cropland ground water NO3– variability. Cropland ground water with an average DO of <3 mg L–1 tended to have <4 mg L–1 of NO3– with a water table depth (WTD) of ≤1 m. Water table depth and DO explained 83% of ground water NO3– variability among buffers. More poorly drained buffers had low ground water NO3– and DO, a shallow WTD, and higher ground water ammonium and soil organic matter. Chloride patterns indicated that dilution was minor in most buffers, suggesting that denitrification losses were important. Soil-landscape factors strongly influenced NO3– behavior and suggest the importance of accurately characterizing soil variability along cropland-riparian zones.

Abbreviations: DO, dissolved oxygen • MWD, moderately well drained • NH4+, ammonium • NO3–, nitrate • OC, organic carbon • OM, organic matter • PD, poorly drained • SPD, somewhat poorly drained • VPD, very poorly drained • WTD, water table depth


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
NITROGEN (N) management in relation to ground water quality and farm profitability is a concern in the USA and many other countries. Nitrate (NO3–) is the most common ground water nutrient in the USA, and concentrations in ground and surface waters are used as critical water quality indicators (Keeney and Follett, 1991). Ground water with NO3– concentrations above the USEPA's maximum contaminant level of 10 mg L–1 are considered unfit for human consumption. Excess NO3– in streams and rivers also contributes to coastal eutrophication. Mueller et al. (1995) estimated that up to 26% of wells near intensive agricultural areas of the USA were contaminated with NO3–. Contamination of ground water with NO3– is associated with the presence of coarse textured soils in urban and intensive agricultural landscapes (Nolan and Stoner, 2000).

Any downward movement of water through the profile can cause NO3– to leach, with the magnitude being proportional to the concentration and water flux (Pierzynski et al., 2005). Nitrate is removed from the soil solution primarily by plant and/or microbial uptake, leaching, and denitrification. Research has shown that small increases in N application rates beyond corn needs can result in high subsurface drainage water NO3– concentrations. For example, Sogbedji et al. (2000) reported that increasing corn N rates from 100 to 134 kg N ha–1 yr–1 nearly doubled average flow-weighted tile drainage water NO3– concentrations for loamy sand and clay loam soils. The authors suggested that 100 kg N ha–1 may be a critical N application rate above which fertilizer N efficiency drops rapidly and the potential for NO3– leaching increases.

Crop fields adjoining riparian areas can contribute NO3– to surface waters via subsurface flow pathways that discharge to streams (Spruill, 2000). Research has demonstrated that a majority of stream discharge in humid regions is typically from soil and ground water inputs (Rice and Hornberger, 1998; Burns et al., 2001; McHale et al., 2002). Much of this water is often presumed to have passed through riparian zones, but the actual flow paths of ground water determine the degree of interaction between NO3––enriched ground water and riparian soils (Hill, 1996; Puckett et al., 2002). Riparian areas across a range of soils and landscapes have been shown to facilitate NO3– reductions in shallow ground water (Haycock and Pinay, 1993; Hill, 1996; Lowrance et al., 1997; Clément et al., 2003). The reduction of NO3– to gaseous forms (NO, N2O, and N2) via denitrifying bacteria, plant and/or microbial uptake of NO3–, and dilution from ground water can decrease NO3– concentrations in riparian zones (Altman and Parizek, 1995; Addy et al., 1999; Clausen et al., 2000; Clément et al., 2002; Kellogg et al., 2005). Denitrification losses in riparian zones are thought to be primarily limited by drainage and organic carbon (OC) availability (Groffman et al., 1992; Lowrance, 1992). Plant uptake of N in forest buffers has been reported (Peterjohn and Correll, 1984), while others have emphasized the role of vegetation in providing readily available OC to fuel denitrification (Groffman et al., 1992; Lowrance, 1992; Haycock and Pinay, 1993; Hanson et al., 1994; Kellogg et al., 2005). Denitrification along deeper ground water flow paths can also occur provided that low redox potentials, sufficient OC, and/or other suitable electron donors are present (Korom, 1992; Nolan, 1999). In general, research has demonstrated NO3– removal in riparian zones is dependent on hydrogeologic setting and soil characteristics, which affect ground water flow path chemistry (Lowrance et al., 1997; Vidon and Hill 2004a, 2004c). In some settings, ground water can flow beneath riparian zones resulting in little opportunity for NO3– removal by surface processes (Puckett et al., 2002; Puckett, 2004). Studies suggest that shallow water tables (e.g., poor drainage) increase the likelihood of denitrification and plant uptake in riparian areas (Simmons et al., 1992; Hanson et al., 1994; Burt et al., 1999; Gold et al., 2001; Clément et al., 2002).

Soil drainage class strongly influences N cycling in agricultural, forest, and riparian soils (Groffman and Tiedje, 1989; Burt et al., 1999; Addy et al., 1999; Briggs et al., 2000; Moiser et al., 2002; van Es et al., 2002). In general, N leaching potential increases as soils transition from poorly drained to well drained, whereas denitrification potential decreases. The soil-landscape model has been used extensively in soil mapping, drainage class prediction, site productivity, and other soil-site relationships (Briggs and Lemin, 1994; Campling et al., 2002; Arnold, 2006; Grunwald, 2006). Variations in drainage class and concentrations of electron donors in shallow ground water can follow strong patterns from upland to riparian areas (Lowrance et al., 1997). Many studies of riparian N dynamics have included limited measurements in uplands, yet these areas can have widely differing N fluxes to riparian zones (Vidon and Hill, 2004a, 2004b, 2004c). Few studies in the Northeast have examined relationships among shallow ground water NO3– and factors such as N inputs, soil series, drainage class, and dissolved oxygen (DO) concentrations among cropland-riparian areas. In addition, there is little field-based information on NO3– leaching to shallow ground water for major New York field crops, while the potential for NO3– reduction by adjoining riparian buffers is also largely unknown.

A study was initiated in 2002 to examine soil and ground water N dynamics among cropland and riparian buffers at two central New York agricultural sites. We hypothesized that cropland N inputs and soil-landscape variability would jointly affect NO3– leaching and transformations in riparian ground water. The objectives of the study were to: (i) quantify inorganic N forms in soil and shallow ground water among paired cropland-riparian buffer plots, (ii) evaluate the relative effectiveness of recently established buffers and established riparian forest buffers on NO3– attenuation in shallow ground water flow, and (iii) examine relationships among N inputs, soil series, drainage class/water table depth, and DO and NO3– concentrations.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Areas
The study sites were located at two farms in adjacent northward draining valleys in Onondaga County, New York. Paired cropland-buffer plots were established along portions of Spafford Creek (42°82' N; 76°23' W) and Onondaga Creek (42°85' N; 76°13' W) in 2002 (Fig. 1). Glaciolacustrine, alluvial, and outwash deposits overlie Devonian-age shale bedrock in both valleys. The valley bottoms have shallow, unconfined aquifers of variable thickness that discharge to the streams (Winkley, 1989). The moderately well (MWD) and well drained (WD) soils formed in gravelly and sandy glacial outwash from relic alluvial fans, terraces, deltas, and kames (Hutton and Rice, 1977). These soils include the Palmyra (fine-loamy over sandy or sandy-skeletal, mixed, active, mesic Glossic Hapludalfs), Lansing (fine-loamy, mixed, active, mesic Glossic Hapludalfs), Howard (loamy-skeletal, mixed, active, mesic Glossic Hapludalfs), and Phelps (fine-loamy over sandy or sandy-skeletal, mixed, active, mesic Glossaquic Hapludalfs) series. Finer-textured glaciolacustrine and alluvial soils that are MWD to somewhat poorly drained (SPD) include the Collamer (fine-silty, mixed, active, mesic Glossaquic Hapludalfs), Hamlin (coarse-silty, mixed, active, mesic, Dystric Fluventic Eutrudepts), and Teel series (coarse-silty, mixed, active, mesic Fluvaquentic Eutrudepts). The more imperfectly drained silty and clayey glaciolacustrine soils include the SPD Rhinebeck (fine, illitic, mesic Aeric Endoaqualfs) and the poorly drained (PD)/very poorly drained (VPD) Wayland series (fine-silty, mixed, active, nonacid, mesic Fluvaquentic Endoaquepts).


Figure 1
View larger version (121K):
[in this window]
[in a new window]

  		Fig. 1. Digital orthophotograph of the (A) Spafford Creek and (B) Onondaga Creek research sites in Onondaga County, NY showing paired cropland-riparian plots and monitoring well locations. Established riparian forest buffer plots at Spafford Creek (SB9, SB10) were located approximately 300 m north of the downstream sampling location (not shown on map). Soil series boundaries represent the original soil survey (1:20000) mapping of Hutton and Rice (1977) from the SSURGO database (ChB = Collamer; Hb = Hamlin, LsC = Lansing, PgB = Palmyra, PpB = Phelps, Rh = Rhinebeck, Te = Teel, and Wn = Wayland).

 
Fields are managed in corn silage (Zea mays L.), alfalfa (Medicago sativa), and/or permanent hay (Phalaris arundinacea) in various stages of rotation (Table 1). Corn fields receive surface-applied liquid dairy manure in the spring at rates of about 70 to 100 Mg ha–1 (wet basis). Manure is generally incorporated within 1 to 4 d after application. All corn N needs are applied at or before planting through a combination of manure, broadcast fertilizer, and banded N at planting. Total N application on corn fields was in the range of 100 to 140 kg N ha–1 after including N credits from previous sod and manure.


View this table:
[in this window]
[in a new window]

  		Table 1. Soil series, drainage class, buffer vegetation, and field crops grown during 2002 to 2005 for paired cropland-buffer plots at the Spafford and Onondaga Creek sites.

 
Riparian Buffers
A cool season grass riparian buffer consisting of Lolium multiforum, Phalaris arundinacea, Pleum pratense, and Dactylis glomerata was planted along the east and west sides of Spafford Creek in August 2002. Ten buffer strip plots (9 x 50 m) were then established along both sides of the stream, with plots spaced 100 to 125 m apart on the same stream side (Fig. 1). In May 2003 existing grass (Phalaris arundinacea) buffers at the Onondaga Creek site were widened to 9 m, and six buffer strip plots (9 x 50 m) were then established (Fig. 1). All recently established grass buffers were then randomly assigned to one of two vegetation treatments consisting of grass alone (n = 6) or in combination with willow (n = 6). Salix discolor cuttings (variety S365) from the SUNY-ESF Willow Biomass Program were planted in June 2003 following moldboard plowing, disking, and treatment with residual herbicide. Willows were planted in two double rows (0.6 x 0.6 m within double rows, and 0.80 m between double rows) along the stream bank edge for a distance of 50 m. Above ground biomass for willow plots was estimated at the end of the 2005 season. Briefly, 30 stems were selected from willow trees at each site, measured, dried, and weighed. Regression equations were used to estimate above ground biomass as a function of stem diameter (T.A. Volk, personal communication, 2005). Published values of N concentration (Tharakan et al., 2003) for Salix discolor were used to estimate N removal. Two buffer plots in existing riparian forest buffers at each site were also established. Forest riparian plots were about 10 m in width, located on PD/VPD hydric soils, and occupied by Salix nigra, Fraxinus pennsylvanica, and Acer rubrum.

Soil Water and Ground Water Sampling
Soil solution was sampled at 25- and 50-cm depths with porous cup (i.d. = 3.5 cm) tension lysimeters (Soil Moisture Equipment Corp., Santa Barbara, CA). Sampling portals were excavated with 5-cm diam. augers at slight angles to minimize potential flow along the lysimeter-soil interface (Mitchell et al., 2001). A paste made from native horizons was used to form a seal between the soil matrix and porous cups. One 25-cm and one 50-cm lysimeter was installed per plot. Buffer lysimeters were installed in April each season to within 1 m of ground water wells. Cropland lysimeters were installed within 1 m of cropland wells after corn was planted in late May. Lysimeters were placed under 50 kPa of tension for 24 to 48 h after rainfall events. Crop fields at the Spafford Creek site are tile drained (about 0.70- to 1.0-m depths). Grab samples of drainage water were taken from the two main outlets on the east side of the stream (outlets near the SB1 and SB4 buffers) that drain approximately 6 ha of cropland, in addition to the main outlet on the west side (>10 ha drainage area) of the stream (near the SB7 buffer). Stream water was sampled upstream of tile outlets and at a former gauging station approximately 1200 m downstream.

Thirty-two shallow ground water monitoring wells were installed during July 2002 and July 2003. Cropland sampling plots were established directly in crop fields, approximately 15 m outside buffer edges. Well transects along stream sides were spaced about 100 to 125 m apart (Fig. 1). Bore holes were completed with an ATV-mounted drill (Giddings Machine Co., Windsor, CO), a hollow stem auger drill (Parratt-Wolff Environmental and Geotechnical Drilling Services, E. Syracuse, NY), and/or by hand with a bucket auger. Completed well depths ranged from about 1.0 to 3.5 m below the soil surface. Bore holes were backfilled with sand and sealed with 15 cm of bentonite. Wells were developed by repeatedly removing ground water for several weeks after installation. Well elevations were surveyed to a common datum in November 2003. Ground water samples were collected during the field season approximately monthly during August 2003 to November 2005. For each event, several well volumes were removed and ground water was allowed to recharge before samples were extracted with a peristaltic pump. Dissolved oxygen concentrations were measured following well recharge with a portable DO meter (Cole-Parmer Instrument Co., Vernon Hills, IL) by carefully lowering the probe to a depth of about 0.5 m below the water table.

Hydrology
Rainfall and air temperature were monitored during 2004 to 2005 with a portable weather station (Spectrum Technologies, Inc., Plainfield, IL). Continuous soil water content of two grass-willow buffer plots (SB1 and OB1) was measured with capacitance probes connected to data loggers (Decagon Devices, Inc., Pullman, WA). Two soil moisture probes were placed in the Ap horizon and three in B horizons by inserting probes into profile walls. Water table elevations in the SC1 and SB1 wells were measured continuously from July to November 2005 using submersible pressure transducers with data loggers (Heron Instruments, Inc., Burlington, Ontario, Canada). Tile drainage water fluxes were monitored during 2005. Tile outlets near the SB7 and SB1 plots were fitted with mini-weirs calibrated to provide a direct measurement of discharge (Thel-Mar Co., Brevard, NC). Water flux for the outlet near SB4 was estimated by the time to fill a known volume. Spafford Creek discharge was calculated by multiplying stream velocity by width and stream stage.

Shallow ground water flux estimates for the upper 1 m of the saturated zone were estimated using Darcy's law (Vidon and Hill, 2004a):

Formula 1[1]
Q is the ground water flux (m3 d–1), Ks is saturated hydraulic conductivity (m d–1), dh/dl is the change in elevation head along the flow direction distance (l) between a given well pair, and A is the saturated unit area thickness (m2). Water table depths were measured in the field with an electronic water level meter (Solinst Canada, Ltd., Georgetown, Ontario, Canada). Saturated area for each well was estimated by multiplying the saturated depth by a width of 1 m to enable comparisons among wells (Lowrance et al., 1984). Saturated hydraulic conductivity was estimated by the Hvorslev slug test method. One transect of nested piezometers (i.d. = 1.9 cm, 30-cm screen length) was established through each of two grass-willow buffers (OB1 and SB1) at the near-stream zone to assess vertical ground water flow and discharge/recharge at the near-stream area. Piezometer nests (one deep and one shallow) were installed on the outside of willow strips and on the inside of the two buffers at the near-stream zone.

Soil and Water Chemical Characterization
Ten random soil cores (0 to 30 cm) were taken from each plot in July 2004. Available P and cations were determined by sodium acetate (Morgan's solution) extraction (0.72 N NaOAc + 0.52 N CH3COOH) following Wolf and Beegle (1995). Filtered extracts were analyzed for P and cations on a PerkinElmer Optima 3300DV inductively coupled plasma-optical emission spectrometer (ICP–OES) (PerkinElmer Corporation, Norwalk, CT) using standard techniques. Soil samples were also analyzed for pH, organic matter by loss on ignition, and total N by the Kjeldahl procedure after Bickelhaupt and White (1982).

Solution samples were collected in polyethylene bottles and immediately placed on ice following collection in the field. Samples were taken back to the lab after each sampling event, stored at 5°C, and filtered (0.45 µm) within 24 h. Samples were analyzed for NO3– + nitrite N (NO2–) and ammonium N (NH4+) using a flow injection autoanalyzer (BRAN+LUEBBE, Roselle, IL). Nitrate plus nitrite (NO2–) concentrations were determined by the hydrazine sulfate-sulfanilamide method with BRAN+LUEBBE Method No. US-696F-82W. Since NO2– was always negligible, the sum of NO3– plus NO2– is referred to as NO3– in this paper. Ammonium N was analyzed by the indophenol blue method with BRAN+LUEBBE Method US-696D-82X. Ground water pH was periodically measured electrometrically in the lab (Accumet Research AR20, Fisher Scientific, Pittsburgh, PA). Ground water chloride (Cl–) concentrations were measured by a flow injection autoanalyzer using BRAN+LUEBBE Method US-696G-82W.

Statistical Analysis
A general linear modeling (GLM) approach was used for hypothesis testing and variance explanation (SAS Institute, 1999). A partially balanced complete block (by site) analysis of variance (ANOVA) was used to test the null hypothesis of equivalent ground water NO3– concentrations among cropland (hay or corn), grass (n = 6), grass-willow (n = 6), and riparian forest buffer plots (n = 4). Least square means were separated by a priori linear contrasts. Soil solution N was analyzed using ANOVA with lysimeter depth (25 and 50 cm) as an additional term. Linear associations among variables were assessed with Pearson correlation coefficients or the Spearman rank correlation procedure for highly skewed variables. Linear and stepwise multiple linear regression were used to examine relationships among soil-landscape factors and ground water NO3–.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
General Soil Properties
The soils at the sites were characterized by a range of drainage classes and fertility (Table 1, 2). Hydric soils (Wayland series) occupied all the riparian forest buffer plots and two grass-willow plots at the Onondaga site. With the exception of the OC2 corn field plot, soil OM was highest in riparian forest buffers. The OC2 location was characterized by a heavy manure application history, high soil P, and anomalously high soil OM compared to other plots.


View this table:
[in this window]
[in a new window]

  		Table 2. Soil pH, NaOAc-extractable P, soil organic matter (OM), total soil N, and soil texture for field plots (0–30 cm) at the Spafford and Onondaga Creek sites.

 
Hydrology
Departures from long-term average monthly rainfall (summed from 1 April to 30 November) for 2003, 2004, and 2005 were –4.25, 14.68, and 4.75 cm, respectively (recorded at Hancock International Airport in Syracuse, NY). There were also large early season rainfall differences (1 May to 31 July) over 2003 to 2005, with departures from long-term average totals of 0.70, 15.28, and –9.90 cm, respectively. In 2005, changes in soil moisture (SB1), ground water elevation heads (SC1–SB1), stream discharge, and tile drainage water fluxes at Spafford Creek followed similar temporal patterns indicating linkage between vadose zone and shallow ground water fluxes (Fig. 2). Tile drainage fluxes were highly correlated (0.73 ≤ r ≤ 0.89, p < 0.0001) among tiles over 2005, and were highly correlated (0.83 ≤ r ≤ 0.89, p < 0.0001) with Spafford Creek discharge. The sum of the average water fluxes from the three tile outlets represented <0.33% of average stream water discharge for the period measured in 2005.


Figure 2
View larger version (30K):
[in this window]
[in a new window]

  		Fig. 2. (A) Hourly rainfall, (B) soil water content in the Ap and B horizon for SB1, and (C) tile drainage and stream water discharge during the 2005 field season at Spafford Creek. Soil water content values are means for Ap (n = 2) and B (n = 3) horizons.

 
Cropland ground water elevation heads were greater than buffer elevation heads for each cropland-buffer well pair in 2004 and 2005. Potentiometric heads in the OB1 and SB1 buffers indicated slight upward flow components (e.g., deeper piezometers had higher pressures) at the near-stream area in 2004 (Fig. 3) and 2005 (data not shown). Ground water elevation heads for the SC1-SB1 well pair were highly correlated (r = 0.92, p < 0.0001) over 2005 (Fig. 3). A simple conceptual model of shallow ground water flow for the sites was created based on three-point ground water flow analyses of monitoring well heads and piezometric heads in the near-stream zone taken during 2004 and 2005. These data suggested overall shallow ground water flow from cropland to buffers at slightly oblique angles relative to stream flow (Fig. 4).


Figure 3
View larger version (36K):
[in this window]
[in a new window]

  		Fig. 3. (A) Potentiometric heads in the SB1 and OB1 grass-willow riparian buffers during 2004 and (B) changes in relative ground water elevation for the SC1-SB1 well pair during 2005 at Spafford Creek. Dashed lines are potentiometric heads in the deeper piezometer for each pair. The first letter indicates site (S = Spafford, O = Onondaga), the third letter indicates relative depth (S = shallow, D = deep), and numbers refer to location (1 = outside of willow strip, 2 = near-stream zone).

 

Figure 4
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 4. Conceptual hydrologic model of shallow ground water flow in the surficial, unconfined aquifers at the sites. The solid arrow on the soil surface depicts predominant horizontal ground water flow direction in upper portions of the saturated zone. The solid arrow within the aquifer depicts slight upward flow components (e.g., discharge) at the near-stream area. The dashed arrow represents potential inputs to stream discharge from deeper, progressively older ground water flow paths. Ground water flow lines are based on three-point ground water flow analyses and potentiometric heads measured by nested piezometers in the near-stream zone during 2004 to 2005. Well depths and water table depth depict average site conditions. Note piezometers and lysimeters are not pictured.

 
Soil Water Nitrate and Ammonium Concentrations
Dry conditions during 2003 produced few soil solution samples and did not permit a statistical analysis of NO3–. Soil solution NO3– concentrations in some grass-willow plots were >20 mg L–1 and similar to concentrations in corn fields in 2003 (data not shown). In 2004 and 2005, corn field plot NO3– concentrations were significantly higher than hay and buffer plots for each event (Table 3). Nitrate peaked for corn and hay plots in mid-June or early July. Mean 25-cm depth corn plot NO3– was consistently higher than 50-cm depth NO3– and significantly greater for 9 July 2004, 30 July 2004, and 16 June 2005. Soil solution 25-cm depth NO3– was significantly correlated (p ≤ 0.01) with 50 cm in cropland for each sampling, with no correlation between 25- and 50-cm depth NO3– in buffer plots. Drier conditions in 2005 resulted in fewer data sets, but trends were similar to 2004 (Table 3).


View this table:
[in this window]
[in a new window]

  		Table 3. Mean 25- and 50-cm depth soil solution nitrate (NO3–) concentrations for cropland and riparian buffer plots for the 2004 to 2005 seasons.

 
Average tile drainage water NO3– over 2003 to 2005 (n = 25) for outlets near SB1, SB4, and SB7 plots was 5.2 (SD = 1.8), 5.0 (SD = 2.2), and 5.7 mg L–1 (SD = 1.7), respectively. Rainfall in May and early June was associated with the largest NO3– increases, whereas rainfall during mid-June to late August had little impact on drainage water NO3– each year (data not shown). Average stream water NO3– was about fivefold lower than tile drainage water and was similar at the upstream (1.0 mg L–1; SD = 0.4) and downstream location (1.2 mg L–1; SD = 0.4). Tile drainage water NO3– concentrations for each outlet were weakly correlated with stream water NO3– (0.50 ≤ Spearman r ≤ 0.67, p ≤ 0.05) over 2003 to 2005.

Soil solution NH4+ concentrations were more variable than NO3– and tended to be higher for 50-cm depth forest buffer plots (data not shown). Mean 25- and 50-cm depth soil solution NH4+ over 2004 to 2005 for all plots was 0.21 mg L–1 (SD = 0.35) and 0.30 (SD = 0.52), respectively. Average 50-cm depth NH4+ in forested plots (3.18 mg L–1) was almost two orders of magnitude greater than other plots (range = 0.04 to 0.10 mg L–1) in September 2003, and about threefold greater for October 2003 (0.65 mg L–1) compared to other plots (0.01 ≤ NH4+ ≤ 0.21). Fifty cm NH4+ concentrations in forest buffers were also highest (0.01 ≤ p ≤ 0.29) in May 2004 (2.4 mg L–1), June 2004 (2.1 mg L–1), and June 2005 (0.92 mg L–1).

Shallow Ground Water Nitrate and Ammonium
Ground water NO3– was consistently highest beneath corn fields throughout the study (Table 4). Ground water NO3– concentrations for corn field plots were not significantly different from hay plots during 2003 to 2004. In 2005, corn field plot ground water NO3– was significantly higher than hay and buffers for five of six sampling events. Nitrate concentrations in corn plots showed some temporal variation. In 2003, concentrations increased from September to October after above average rainfall. Increased NO3– leaching to ground water also occurred in July 2004 following a period of rainfall. In 2005, mean ground water NO3– in corn plots reached a maximum of 7.1 mg L–1 after significant rainfall around 16 June, which was followed by a 50% decrease in July, and an 80% increase on 3 September (Table 4). Ground water NH4+ tended to follow opposite trends of NO3– and was generally higher in forest buffers. Ground water NH4+ in 2003 was below detection limits for corn field and grass buffer plots and significantly higher in forest buffers. The highest mean NH4+ (about 4.4 mg L–1) for the study occurred in riparian forest buffer plots in August and September 2003 (Table 4).


View this table:
[in this window]
[in a new window]

  		Table 4. Least square means for ground water nitrate (NO3–) and ammonium (NH4+) for cropland and riparian buffer plots for the 2003 to 2005 seasons.

 
Soil-Landscape Influences on Ground Water Nitrate and Ammonium
Water table depth (WTD) was positively correlated with ground water NO3– among cropland and buffers for each 2004 and 2005 sampling event (p ≤ 0.01) and for study averages (r = 0.68, p < 0.0001) (Fig. 5). Average WTD and NO3– were also positively correlated within cropland (r = 0.61, p = 0.01) and buffers (Spearman r = 0.71, p = 0.003), and were significantly correlated (p ≤ 0.01) for each sampling event. The OB4 plot was not included for these relationships, as this buffer is comprised of nonnative soil from a previous conservation project. Ground water NO3– was significantly correlated (p ≤ 0.001) with DO for each sampling and for study averages (r = 0.75, p < 0.0001) among all plots, within cropland (r = 0.65, p = 0.001), and buffers (Spearman r = 0.84, p < 0.0001) (Fig. 5).


Figure 5
View larger version (30K):
[in this window]
[in a new window]

  		Fig. 5. Relationships between (A) study average ground water nitrate (NO3–) and average water table depth (WTD), (B) average ground water NO3– and average ground water dissolved oxygen (DO), (C) average ground water DO and average WTD, and (D) soil organic matter (OM) and average WTD for riparian soils.

 
There was a weak correlation (Spearman r = 0.57, p = 0.02) between mean 50-cm depth NO3– and ground water NO3–. Stepwise multiple linear regression selected (p ≤ 0.05) DO and 50-cm depth NO3– as significant predictors of study average cropland ground water NO3–, accounting for 64% of its variation (Fig. 6). Including WTD in the model did not significantly increase the proportion of variance explained (67%). Dissolved oxygen and WTD explained 83% of riparian ground water NO3– variation, but the distribution of model residuals was skewed. Study average WTD and DO were correlated (r = 0.52, p = 0.003) among all wells (Fig. 5), with weaker correlations within cropland (r = 0.54, p = 0.04) and buffers (r = 0.45, p = 0.08). In riparian soils, there was a significant negative correlation (r = –0.64, p = 0.01) between WTD and soil OM (Fig. 5). Averages for WTD, ground water NO3–, NH4+, and DO among plots generally support the correlation and regression results among these variables (Fig. 7). Ground water DO and NO3– were lowest in forest buffers and highest in corn field plots, whereas NH4+ was highest in forest buffers. Ammonium was weakly negatively correlated with NO3– (Spearman r = –0.62, p = 0.0002). There were also weak negative correlations between NH4+ and DO (Spearman r = –0.58, p = 0.0005), and NH4+ and WTD (Spearman r = –0.38, p = 0.03).


Figure 6
View larger version (18K):
[in this window]
[in a new window]

  		Fig. 6. Relationship between observed average cropland ground water nitrate (NO3–) and average cropland ground water NO3– predicted by multiple linear regression with study average 50-cm depth soil solution NO3– and ground water dissolved oxygen as independent variables. Bands represent 95% confidence intervals for the regression line.

 

Figure 7
View larger version (26K):
[in this window]
[in a new window]

  		Fig. 7. Study averages for shallow ground water nitrate (n = 14), ammonium (n = 14), dissolved oxygen (n = 9), and water table depth (n = 14) among cropland and buffer plots over 2004 to 2005. Error bars are estimated standard errors for within plot means.

 
Average ground water NO3– fluxes were higher in cropland (mean = 35, SD = 79) compared to buffers (mean = 12, SD = 27), with high variability in space and time (Table 5). Hydraulic conductivity estimates ranged over three orders of magnitude, contributing to the high variability in ground water fluxes. Nitrate reduction by buffers depended on drainage class. Buffers with WD soils (SB7 and SB8) had large NO3– fluxes to the near-stream zone (Table 5). Apparent dilution from discharging ground water occurred in the SB1 (mean C1– dilution = 75%, SD = 19), SB2 (87%, SD = 11), and SB7 (58%, SD = 32) buffers (Table 5), while other buffers showed minor decreases or increases in Cl– concentrations relative to cropland.


View this table:
[in this window]
[in a new window]

  		Table 5. Average ground water nitrate (NO3–), ammonium (NH4+), dissolved oxygen (DO), water table depth (WTD), ground water NO3– flux, and chloride (Cl–) dilution for paired cropland-buffer plots at the Onondaga (O) and Spafford Creek (S) sites.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Soil Series and Drainage Class
Drainage class and parent material primarily differentiate soil series, reflecting variations in parent material deposition, WTD, and microtopography (Hammer, 1998). Drainage class expresses a range in relative wetness and is indexed by the depth below the soil surface to redoximporphic features, which reflects the long-term seasonally high water table. Drainage class was expressed as a continuous variable by the field-measured WTD. Since relatively long well screens (1.5 m) were used, each WTD was also the upper boundary of the shallow unconfined aquifer. Seasonally high WTD trends directly confirmed the drainage classes described in the soil survey, despite the presence of tile drainage at Spafford Creek.

Soil Solution Nitrate Trends
Cropland soil solution NO3– was up to three orders of magnitude greater than buffer concentrations and depended on the quantity of N applied (e.g., corn vs. hay). This finding agrees with other studies showing greater NO3– availability and leaching with increasing N application (Jemison and Fox, 1994; Randall and Iragavarapu, 1995; Brye et al., 2001; van Es et al., 2002). In 2004 and 2005, corn plot NO3– concentrations at 25 and 50 cm peaked after the first or second rainfall early in the growing season. Fertilizer and manure were applied to corn fields during May and/or early June each year. Fertilizer N can be rapidly transformed to NO3– in the soil solution, and likely contributed to the higher concentrations in late spring and early summer before maximum uptake by corn. The steady decline in soil solution NO3– over the growing season largely reflects uptake by crops and vegetation, leaching, and probably some denitrification and microbial immobilization. Soil NO3– differences among 25- and 50-cm depths in corn field plots were more pronounced earlier in the growing season, with smaller differences as the growing season progressed. Long-term agricultural activity has homogenized upper horizons at these sites, and the alluvial soils have little inherent differentiation between Ap and B horizons. These factors probably contributed to the lack of NO3– differences among depths.

High soil solution NO3–concentrations in 2003 for grass-willow plots were attributed to sod plow-down before planting willows. Moldboard plowing cool season grasses can release up to 70 kg ha–1 of available N during the plow-down year (Ketterings et al., 2003). Mean 25-cm depth NO3– in grass-willow buffers was about 3 mg L–1 in April 2004, but concentrations were equivalent to grass by June 2004. Average N removal in above ground willow biomass at the Onondaga site (176 kg N ha–1) and Spafford Creek (52 kg N ha–1) after about 2.5 yr of growth varied considerably due to differences in yield (38.5 vs. 11.3 metric t ha–1). Willows were presumably effective scavengers of soil solution NO3–. Rapid growth and high N uptake by Salix discolor and other willows are well established (Tharakan et al., 2003).

Cropland Ground Water Nitrate Patterns
Soil solution NO3– was a good indicator of N inputs and availability among cropland plots, but was not independently useful in predicting NO3– leaching to shallow ground water. Ground water NO3– concentrations were more strongly related to soil drainage class and DO. Correlations between cropland soil NO3– and associated ground water NO3– were weak for individual sampling events and study averages. Sogbedji et al. (2000) stated the PSNT-type measures of N availability can effectively identify N-sufficient maize sites, but have not been directly related to water quality measurements (e.g., shallow ground water NO3–). The lack of correlation between soil NO3– and ground water NO3– in this study is partly related to the variable distances and biochemical transformations that occur between the rooting zone and the water table. These ‘intermediate vadose zone’ areas, particularly near the water table, can support high denitrification rates (Pionke and Lowrance, 1991). For example, Puckett (2004) suggested that denitrification in upland ground water can occur in any setting where flow paths encounter reducing conditions.

The relationship between drainage capacity and NO3– leaching potential is used to categorize NO3– leaching potential for NY agricultural soils (Ketterings et al., 2003). Well drained soils tend to have greater leaching potentials and are generally at greater distances above the water table due to landscape position relative to more PD soils. The WD soils at our study sites formed in outwash and had seasonally high WTD of about 0.6 m. Ground water NO3– was consistently higher in outwash soils (SC1, SC7, SC8, OC1, OC3–OC6) relative to more PD alluvial and glaciolacustrine soils (SC2–SC6, SC9, S10) receiving similar N inputs. Cropland plots with an average WTD of ≥1 m tended to have ≥2 mg L–1 NO3. Simmons et al. (1992) showed a significant correlation between WTD and ground water NO3– in MWD upland soils adjoining forest buffers. Nolan (2001) found a significant correlation between measured WTD (monitoring wells) and NO3– in recently recharged ground water from mixed land use areas of the Southeast. These studies concluded that a shallow WTD would favor greater denitrification and lower leaching potentials. Results from our sites support this hypothesis and showed that more imperfectly drained cropland soils had significantly lower ground water NO3–, suggesting a greater tendency for denitrification losses from these zones.

The stepwise regression selected 50-cm depth NO3– as a significant predictor of ground water NO3– in combination with DO. The fact that WTD was not selected suggests the presence of redundant information in the full regression model due to the significant correlation between WTD and DO. The covariance between WTD and DO indicates that DO, and presumably redox potential, decreased as soils became more PD. Oxygen concentration is a critical factor affecting denitrifcation in agricultural soils, and lower DO typically relates to higher soil water contents (Meisinger and Randall, 1991). Nolan (1999) reported that ground water with <0.05 mg L–1 NO3– had associated DO of ≤2.1 mg L–1. Study averages for cropland ground water at our sites showed that DO concentrations <3 mg L–1 tended to have <4 mg L–1 NO3–, and a WTD greater than or equal to about 1 m. More PD field areas would also increase interaction between ground water and upper profile areas, where OC availability would tend to be higher. It is possible that moisture was the limiting factor for potential denitrification in these soils because OC availability is generally not considered limiting in fields receiving manure and/or crop debris (Meisinger and Randall, 1991; Brye et al., 2001). The relationships among WTD, DO, and NO3– further suggest that ground water NO3– concentrations in cropland were constrained by drainage class and associated thermodynamic constraints on N transformations.

Riparian Buffer Ground Water Nitrate and Ammonium Patterns
Studies have shown that soil drainage class and site topography can affect NO3– removal in riparian areas (Hanson et al., 1994; Addy et al., 1999; Burt et al., 1999; Clément et al., 2002, 2003; Vidon and Hill, 2004c). The combination of WTD and DO accounted for 80% of average riparian ground water NO3– variability. Where riparian soils were MWD/WD with DO >3 mg L–1, NO3– concentrations and fluxes were high. For example, the SB7 and SB8 (WD/MWD) buffers functioned as continuous sources of NO3–. These buffers also had very low NH4+, higher average DO (3.7 mg L–1), and greater WTD. Buffers on outwash deposits may have limited NO3– removal capacity due to deeper flow paths and the inability to elicit reducing conditions. Burt et al. (1999) showed that some riparian areas along the River Thame near Oxford, England were ineffective at reducing NO3– due to transport of farmland drainage water through gravel lenses beneath the floodplain directly to the stream channel. In contrast to outwash soils, the more imperfectly drained riparian soils along Spafford Creek (SB2, SB3, SB4, SB5, and SB6) had study average NO3–, NH4+, DO, and WTD of 0.01 mg L–1, 0.12 mg L–1, 1.4 mg L–1, and 0.90 m, respectively. These DO values would appear to support denitrification, while the elevated NH4+ also suggests reducing conditions and limitations on nitrification reactions (McClain et al., 1994; Nolan, 1999; Spruill, 2000). Gold et al. (2001) reported an average ground water DO of 9 mg L–1 for MWD soils and 3 mg L–1 in hydric soils, with associated denitrification potentials of <0.1 µg N kg d–1 and 12 to 45 µg N kg d–1, respectively. Vidon and Hill (2004a) showed that riparian zones with <2.1 mg L–1 DO had strongly enriched {delta}15N values indicative of denitrification. In our study, average NO3–, NH4+, DO, and WTD in riparian forest buffers (hydric soils) were 0.10 mg L–1, 0.77 mg L–1, 1.7 mg L–1, and 0.53 m, respectively. These conditions suggest favorable conditions for both plant uptake and denitrification due to increased interaction of ground water with upper soil horizons higher in OC and plant roots. The more PD riparian buffers also had higher OM content, which should increase OC availability relative to more WD buffers. Burt et al. (1999) suggested NO3– removal in floodplain ground water is dependent on the elevation of the water table and that a high WTD is required to maximize denitrification. Clément et al. (2002) studied denitrification patterns among three riparian areas of differing vegetation in Brittany, France and found no differences in denitrification among forest, understory, and grass buffer sites. The authors concluded that site topography and WTD primarily controlled NO3– removal rather than differences in vegetation and/or OC availability. Residence time of shallow ground water can also affect denitrification and the volume of water that can be denitrified (Burt, 2005), but there was no correlation between seepage velocity or hydraulic conductivity and NO3– for buffers in this study. Our results suggest that WTD/drainage class strongly influenced the fate of NO3–, presumably from constraints on redox conditions along ground water flow paths and associated impacts on denitrification.

Puckett (2004) summarized results from thirteen studies of NO3– removal efficiency in riparian zones across the USA and proposed five hydrogeologic factors explaining variability among sites: (i) denitrification in the up-gradient aquifer due to the presence of OC or other electron donors, (ii) long residence times (>50 yr) along ground water flow paths allowing even slow reactions to completely remove NO3–, (iii) dilution of NO3––enriched waters with older water having little NO3–, (iv) bypassing of riparian zones due to extensive tile drains and ditches, and (v) movement of ground water along deep flow paths below reducing zones. We hypothesize that shallow redox zones in the upland aquifer beneath crop fields, largely caused by drainage class variability, resulted in lower ground water NO3–. More PD crop fields would be expected to contribute lower NO3– fluxes to the riparian zone. Therefore, NO3– removal by riparian zones down-gradient from these fields may be considered less critical from a water quality standpoint relative to WD crop fields.

Ground water sampled at our study sites presumably represented relatively recent recharge. The use of shallow wells should have reflected primarily younger ground water flow paths, particularly in cropland. Average hydraulic conductivity and gradients were used to estimate seepage velocity and indicated that ground water could have traveled from cropland to buffers within 2 to 4 mo, which suggests relatively short residence times and young ground water flow paths. However, buffers with piezometers installed did indicate slight upward flow components. Assuming Cl– patterns were reflective of ground water dilution among buffers, dilution does not explain the bulk of the observed NO3– decreases. All but two buffers showed trivial decreases in Cl– for the events measured, which suggests denitrification was important. Chloride increased in several buffers at the sites. Increased Cl– in buffers has been observed in other studies and attributed to inputs from agriculture and/or road salt applications (Anderson, 1993; Puckett et al., 2002; Vidon and Hill, 2004a), which both occurred at the sites. A decrease in the saturated unit-area thickness in riparian areas could have also contributed to increased Cl– concentrations. It is possible that deeper, older flow paths discharging beneath buffers and the near-stream could have contributed Cl– from up-gradient inputs, while NO3– in the same water could have been removed via denitrification during the longer residence time (Puckett et al., 2002). Notwithstanding, average riparian ground water Cl– for each sampling at Spafford Creek was within 2 to 8 mg L–1 of stream water Cl– concentrations, suggesting that the overall riparian ground water chemistry measured was reasonably indicative of ground water discharging to the stream. Limitations with our hydrological characterization, such as the inability to accurately describe three-dimensional ground water flow, coupled with the high spatial variability of hydraulic properties, does not permit an analysis of the contribution and chemistry of potentially deeper ground water paths discharging beneath buffers and the stream.

It is possible that the tile drainage system at the Spafford Creek site caused some portion of ground water NO3– fluxes to bypass buffers, which could have decreased NO3– removal efficiency. If the tile drains significantly lowered the elevation of the surficial aquifer, some of the lower NO3– concentrations in buffers could have been caused by the absence of ground water containing high NO3– concentrations (e.g., NO3– in surficial flow from cropland flowed beneath buffer wells). However, artificial drainage did not change any of the original soil drainage classes for cropland or buffers based on seasonally high WTD. Additionally, drainage water from the three outlets represented >15 ha of cropland, yet the sum of the average fluxes only accounted for <0.33% of average stream discharge in 2005. These data suggest that the tile drainage configuration may not have altered localized shallow ground water flow enough to cause significant bypassing of most riparian zones. In buffers that were MWD/WD and did not have tiles, ground water flow was deeper in the profile, and presumably flowed beneath more biologically active soil zones.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Cropland and riparian buffer N patterns in relation to soil-landscape factors were examined at two agricultural sites in central New York. Soil drainage class/water table depth and shallow ground water DO explained significantly more cropland ground water NO3– variability relative to soil solution NO3–. Well drained and MWD corn fields consistently had the highest concentration and fluxes of ground water NO3–, while concentrations and fluxes for SPD soils receiving comparable N rates were lower. Soil drainage class and ground water DO also constrained ground water NO3– concentrations and flux in buffers. Well drained or MWD riparian buffers were less effective at NO3– removal relative to PD and VPD soils. Results indicate that soil drainage class and ground water redox had strong effects on N leaching to shallow ground water in cropland soils and on N transformations in riparian ground water. Our results showed that soil variability at scales that may be not captured by typical soil surveys (1:15840 to 1:31680) significantly impacted NO3– dynamics among cropland-riparian areas. Such findings have implications for the development of cost-effective nutrient management practices and water quality modeling efforts.


   ACKNOWLEDGMENTS
 
We thank the USEPA for funding, the Snavlin and CoVale dairy farms, Parratt-Wolff Drilling Inc., Tim Volk, Larry Abrahamson, Ken Burns, Don Bickelhaupt, the USDA-NRCS, and the Onondaga County Soil and Water Conservation District. We also acknowledge the reviewers and the Associate Editor for providing helpful comments on the manuscript.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
E. O. Young and R. D. Briggs
Phosphorus Concentrations in Soil and Subsurface Water: A Field Study among Cropland and Riparian Buffers
J. Environ. Qual., January 4, 2008; 37(1): 69 - 78.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Young, E. O.
Right arrow Articles by Briggs, R. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Young, E. O.
Right arrow Articles by Briggs, R. D.
Agricola
Right arrow Articles by Young, E. O.
Right arrow Articles by Briggs, R. D.
Related Collections
Right arrow Field-Scale Studies
Right arrow Nutrient Cycling
Right arrow Ground Water Quality

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome