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


     

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 ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (15)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by de Vos, J.A.
Right arrow Articles by Raats, P.A.C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by de Vos, J.A.
Right arrow Articles by Raats, P.A.C.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by de Vos, J.A.
Right arrow Articles by Raats, P.A.C.
Soil Science Society of America Journal 64:517-527 (2000)
© 2000 Soil Science Society of America

DIVISION S-1-SOIL PHYSICS

Nitrate Leaching in a Tile-Drained Silt Loam Soil

J.A. de Vosa, D. Hesterbergb and P.A.C. Raatsc

a Wageningen Univ. and Research Centre, Res. Inst. for Agrobiology and Soil Fertility, P.O. Box 14, 6700 AA Wageningen, The Netherlands
b Dep. of Soil Sci., Campus Box 7619, North Carolina State Univ., Raleigh, NC 27695-7619 USA
c Wageningen Univ. and Res. Centre, Dep. of Agric., Environ., and Systems Technol., Dreijenplein 4, 6703 HB Wageningen, The Netherlands

j.a.devos{at}alterra.wagur.nl


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
Nitrate (NO3) leaching was studied for a winter leaching period in a layered calcareous silt loam with tile-drains at about 1-m depth and 12-m spacing. Groundwater levels, drain discharge rates, and NO3 concentrations in the drainage water were monitored, and the soil hydraulic characteristics were measured for the different soil layers. The data were interpreted using the two-dimensional water flow and solute transport model SWMS_2D. This model uses Darcy's law for water flow and the convection–dispersion equation for solute transport for both the saturated and unsaturated zones. A nitrogen-production term of 39 kg N ha-1 was used to account for the net N mineralization in the topsoil during the leaching period. The model was calibrated by varying the hydraulic conductivity at saturation (Ks) for the different soil layers, using the measured groundwater level–drain discharge rate relationship as calibration target. Peaks in NO3 concentrations in the drainage water are well explained by the temporal two-dimensional behavior of convective transport. Measured NO3 leaching was 11 kg N ha-1 yr-1 and simulated NO3 leaching was 15 kg N ha-1 yr-1 in the relatively dry winter leaching period 1991–1992. The two-dimensional transport model SWMS_2D is a useful tool to evaluate the relative effects of management practices to reduce N leaching.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and methods
 Results
 Discussion
 Conclusions
 REFERENCES
 
THE N CYCLE IN AGRICULTURAL SOILS determines N availability to crops and potential N losses to the environment, like NO3 leaching, ammonia volatilization, and denitrification. Nitrate leaching to groundwater or surface waters is a problem because NO3 concentrations often exceed acceptable contamination limits (Strebel et al., 1989; EU, 1991; Fletcher, 1991). New methods for improved N management have to be developed to reduce NO3 leaching from agricultural fields (Keeney and Follett, 1991). Understanding the NO3 leaching process is necessary to assess the consequences of new agricultural management practices.

Nitrogen transformations in soil involve biological processes, like mineralization, immobilization, nitrification, and denitrification. De Willigen (1991) compared 14 models describing N transformations in the soil–crop system. The more complex, mechanistic models did not give better predictions than simpler, functional models, but they may be of more use in comparing management practices. The description of the soil biological processes remains the main difficulty in modeling N transformations.

Feyen et al. (1998) reviewed different types of water flow and solute transport models for homogeneous and heterogeneous soils. Solute transport in homogeneous soils is well described by the classical convection–dispersion model. More complex models have to be used to describe soil heterogeneity, such as a two-region model (van Genuchten et al., 1984; Nieber and Misra, 1993) to discriminate between soil matrix and macropores, including an exchange term for the interaction between the two regions. However, the input parameters and the validation of complex models remain a major problem.

Transport phenomena are often analyzed for a one-dimensional situation and focus on the unsaturated zone (Addiscott and Wagenet, 1985). However, water flow (Van Schilfgaarde, 1974; Zaradny and Feddes, 1978) and NO3 transport (Cannell et al., 1984; Harris et al., 1984) to tile-drains have a two-dimensional character and both the saturated and unsaturated zones are involved. This study addresses water flow and NO3 transport to a tile-drain, combining a detailed field experiment with a two-dimensional transport modeling approach. The field experiment was conducted in a tile-drained silt loam soil with a strongly fluctuating water table. The winter period was considered because this is the period with a substantial excess of precipitation. Soil temperatures were low and N transformations are expected to be slow. Our hypothesis was that the temporal behavior of NO3 leaching in a winter leaching period is dominated by the physical transport processes rather then by the N transformations. The objective was to demonstrate that a model describing two-dimensional water flow and NO3 transport can explain the temporal behavior of NO3 leaching to tile drains. The SWMS_2D model (Simunek et al., 1996) was used, which is based on two-dimensional, transient Darcian water flow and the convection–dispersion transport of solute in the combined saturated–unsaturated zones. Macroscopic soil heterogeneity was described by using the soil hydraulic characteristics for different soil layers. A simple N-production term was used to describe the net effect of the N transformations in soil. Field measurements were used to evaluate this modeling concept.


   Materials and methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
Results
 Discussion
 Conclusions
 REFERENCES
 
The field experiment was conducted in the winter leaching period 17 Dec. 1991 to 8 Apr. 1992. To define the agronomical and hydrological system studied, the experimental field is described first. Input for the SWMS_2D model is obtained from the field data. The description of the soil profile and corresponding soil hydraulic characteristics are used to define the different layers in SWMS_2D. Water flow to tile drains has a horizontal and a vertical component, and therefore the hydraulic conductivities at saturation Ks were measured in both directions. Initial conditions for water flow in SWMS_2D are inferred from measured groundwater levels. Hydraulic heads were measured to analyze whether upward or downward seepage occurred, which determines the bottom boundary condition for water flow in SWMS_2D. Measured soil NO3 concentrations are the initial conditions for NO3 transport. Meteorological measurements are used to define the boundary conditions at the soil surface. The general aspects of the SWMS_2D model are described briefly. The emphasis is mainly focused on the special aspects of using SWMS_2D for our particular experimental field. Field measurements of drain discharge rates, NO3 concentrations in the drainage water, and hydrological conditions are used to evaluate the modeling concept.

Experimental Field
Agronomic Characteristics
The experimental field was located at the Dr. H.J. Lovinkhoeve experimental farm at Marknesse in the Noordoostpolder, the Netherlands. The Lovinkhoeve soil is a calcareous silt loam (Brussaard et al., 1988). The experimental field had been under an integrated arable farming system since 1989. This meant a reduction of 25 to 40% in mineral N fertilizer input, resulting in an average application rate of 110 kg N ha-1 yr-1, and the application of more organic fertilizer, like mushroom compost (Lebbink et al., 1994). A 4-yr crop rotation of potatoes (Solanum tuberosum L.)–winter wheat (Triticum aestivum L.)–sugar beet (Beta vulgaris L.)–spring barley (Hordeum vulgare L.) had been practiced on the field for 27 yr. Sugar beet was grown prior to the winter leaching period studied. After harvesting the sugar beet on 25 Sept. 1991, plant tops and leaves were removed from the field to prevent an input of readily mineralizable organic matter.

Tile-Drained Field Plot and Drainage Water Sampling Device
The Lovinkhoeve has subsurface tile drains at about 1-m depth and 12-m spacing. The tiles (i.d. = 5 cm; o.d. = 7.5 cm) are located at the bottom of a 20-cm-wide drain trench, which was refilled after the drains were installed. The field plot was drained by a single tile drain and was isolated from other parts of the field having different fertilization and soil tillage treatments (Fig. 1) . An experimental area of 62.5 by 12 m2 was thus created, assuming that the watershed boundary will be exactly in the middle between the tile drain of our plot and the adjacent drains. The depth of the tile drain ranged from 90 cm at the end of the experimental field (Fig. 1, location M) to 105 cm close to the ditch (Fig. 1, location D).



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 1 Overview of the tile-drained field plot at the Lovinkhoeve. The thick lines indicate a series of piezometers, the thick dots represent groundwater level observation wells. The position of the automated observation well is also shown. The numbers at the top indicate the blocks at different distances from the drain. The roman numerals at the right hand side indicate the different instrument transects. Location M corresponds to the end of the experimental field; location D is the position close to the ditch

 
The drainage water sampling device monitored the drain discharge rate and took flow-proportional subsamples of the drainage water. A modified sampling system was developed which collected the drainage water in a main reservoir (Fig. 2) , inspired by the design of van Ommen et al. (1989). Whenever this main reservoir was filled (365 L), a sample of this water was taken, which corresponded to 0.5 mm of cumulative drain discharge, and the reservoir was emptied. The times at which the main reservoir was completely filled were recorded, so the average drain discharge rate over each time interval could be calculated (de Vos, 1997). The flexible PVC tube and the float in the ditch (Fig. 2) were used to compensate for the fluctuating ditch water level.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 2 Schematic illustration of the drain sampling process. The arrows indicate the water flow direction

 
Soil Profile
The Lovinkhoeve soils have AC profiles and are classified as Typic Fluvaquents (Soil Survey Staff, 1975) or as Calcaric Fluvisols (FAO-UNESCO, 1974). In the Dutch soil classification system, they are calcareous polder vague soils with a loamy surface layer, mapping unit Mn25A (de Bakker and Schelling, 1966; de Bakker, 1979). The top 25-cm layer has a CaCO3 content of 9%, and the particle-size distribution is: 12% sand, 68% silt, and 20% clay (after Van Faassen and Lebbink, 1990). Figure 3 shows other physical and chemical data derived from Kooistra et al. (1989) or measured in this study. Depths to soil horizon boundaries were measured at various locations throughout the field as soil samples were collected for NO3 analysis. As indicated by the standard deviation bars in Fig. 3, horizon boundaries in the experimental field were rather consistent.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 3 The soil profile of the field plot. Clay content, organic matter content (OM), and cation-exchange capacity (CEC) are indicated for each horizon. The error bars give the standard deviation of the horizon depths

 
Nitrate Concentrations in Soil
Soil nitrate contents and volumetric water contents (100-cm3 samples) were measured in the soil profile at 10-cm intervals between 0- and 150-cm depth at the start and at the end of the experiment. In each block with a different distance from the drain (Fig. 1), five soil samples from each depth interval were taken at random positions, on both sides of the drain. Ammonium contents in the soil were low and are neglected.

Hydraulic Characteristics of the Unsaturated Zone
Water retention characteristics were determined using the filter-funnel suction method (Klute, 1986) for the pressure head (h) range -300 < h < 0 cm; the tensiometer method (Boels et al., 1978) for -800 < h < -300 cm; and pressure extractors (Klute, 1986) for h < -800 cm. The crust method was used to measure the hydraulic conductivity -70 < h < 0 cm (Hillel and Gardner, 1970); for h > -70 cm the hot air evaporation method was used (Arya et al., 1975). de Vos et al. (1992, 1994) discussed the data and presented the corresponding analytical functions of van Genuchten (1980) and Mualem (1976) for the different soil layers. These functions are used in the calibration of the water flow part of the SWMS_2D model (Table 1) .


View this table:
[in this window]
[in a new window]
  		Table 1 Parameters in the analytical expressions of van Genuchten (1980) and Mualem (1976) for describing the hydraulic characteristics of the different layers of the Lovinkhoeve soil after the Ks calibration procedure and the adaptation of the water retention characteristic of the 0–25 cm topsoil (de Vos et al., 1999)

 
Hydraulic Conductivity at Saturation
The hydraulic conductivity at saturation Ks was measured by applying a constant water level at the top of an undisturbed soil sample (19-cm diam. and 10-cm height) and maintaining a water layer below the bottom of the sample to guarantee saturation of the whole sample (de Vos, 1997). This improved method resulted in Ks values a factor 2 to 3 larger and smaller variation (<10%) when measurements on the same samples were repeated three times in comparison with the standard method with free drainage at the bottom of the sample (Klute and Dirksen, 1986; NEN-5789, 1991). The Ks data are shown in Fig. 4 for measurements in the vertical and horizontal directions. Variation in Ks between the samples at each depth is one to two orders of magnitude for both directions. However, a decreasing trend in Ks can be observed for the soil layers below 50-cm depth. At 120-cm depth, the soil samples contained pure peat, which had a distinctly lower Ks than all other layers. For the soil layers above 50-cm depth, Ks was found to be in the range 10 < Ks < 100 cm d-1. Within each soil layer, there was no difference in Ks for the vertical and horizontal directions due to a possible microlayering, except for the soil layer at 38- to 58-cm depth, where Ks for the horizontal direction was substantially higher; also higher than Ks at all other depths. This high Ks is attributed to the presence of a sandy layer, of about 3- to 5-cm thickness, in this plot found between 36- to 46-cm depth. The Ks in the vertical direction in the drain trench was larger than in the other parts of the field.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 4 Hydraulic conductivity at saturation measured in the vertical and horizontal directions at different depths in the soil profile and the drain trench

 
Hydrological Field Measurements
Observation wells, PVC tubes that were perforated over the whole length, were installed to 200-cm depth to monitor groundwater levels. Piezometers, PVC tubes that are only perforated at the bottom, were installed at 125-, 150-, 175-, and 200-cm depths to monitor hydraulic heads. The measurement locations in the field were based on expected variations in hydrologic conditions as a function of the distance from the drain and the change in drain depth over the length of the field (Fig. 1). Measurements were made twice a week along Transect 2 and once a week along Transects 1 and 3 (Fig. 1). The groundwater level was also measured automatically every 15-min midway between the tile drain of our field plot and the adjacent drain (Fig. 1).

Meteorological Station
An automated meteorological station was located 75 m from the experimental plot. Global and net radiation, air temperature, relative humidity, precipitation, wind speed, wind direction, soil temperatures, and groundwater level were measured every 5 s, then averaged and stored as data for each 10-min interval.

The SWMS_2D Model for Water Flow and Solute Transport
The SWMS_2D model (Simunek et al., 1994, 1996) simulates Darcian water flow in a two-dimensional flow domain simultaneously in the unsaturated and saturated zones. Solute transport is described using the convection–dispersion approach. The model SWMS_2D solves numerically the partial differential equations using a finite-elements scheme. In the present study SWMS_2D is applied for a winter leaching period for a bare soil only, thus without water uptake by plant roots. The measured NO3 concentration profile on 17 Dec. 1991 is used as the initial condition, including the spatial distribution of NO3 as a function of the distance to the drain. These data are presented in the later subsection Field Measurements.The net effect of N transformations was simulated by a constant NO3 production term of 0.34 kg N ha-1 d-1, homogeneously distributed through the 0–25 cm topsoil. This cumulative production is 39 kg N ha-1 in the entire winter leaching period, which corresponds to the estimate based on the field N balance. The initial groundwater level midway between the drains was taken as representative for the depth of the initial water table. An equilibrium in hydraulic head was assumed with this water table to calculate initial pressure heads.

Flow Domain and Finite Element Grid
Water flow and solute transport were modeled in the flow domain of 6-m width, representing half the drain spacing, and a soil profile depth of 2 m. The bottom of the soil profile was assumed to be impermeable. In the model, the drain was located at 97.5-cm depth, and was described as a half circular hole with the real physical dimensions. The inner wall of the drain was described as a seepage face, implying that the drain is always practically empty. A finite element grid was created (1352 nodes, 2563 elements) with the triangular elements forced to be stratified according to the layered soil profile (de Vos et al., 1999). The distribution of the triangles was generated using an automatic grid generator (Simunek et al., 1996).

Hydraulic Conductivity Close to Saturation
To obtain a smooth transition between K in the unsaturated and saturated zones, the option in the SWMS_2D model is used to define a volumetric water content {theta}k with corresponding pressure head hk, above which the unsaturated hydraulic conductivity varies linearly between Kk and Ks. This procedure reduces numerical problems at nodal points close to saturation. The Mualem (1976) equation for the unsaturated hydraulic conductivity is used for water contents {theta} < {theta}k. A threshold was chosen for the linear part of the hydraulic conductivity characteristics (de Vos et al., 1999). Smaller hk values were tried, but resulted in numerical stability problems. One value of hk for all soil layers was chosen to keep the number of model parameters minimal. The only difference between the hydraulic conductivity parameters of layers 4, 5, 6, 7, and 8 is the value of Ks (Table 1) and the corresponding steepness of K between .

Calibration of the SWMS_2D Model
The calibration of the water flow and solute transport parts of the SWMS_2D model is described in detail by de Vos et al. (1999). A brief description of their findings follows. The Ks values for the different soil layers are used as calibration parameters and the drain discharge rate and groundwater levels are used as calibration targets. An average is chosen for all soil layers as a first estimate, together with measured soil hydraulic characteristics for the unsaturated zone described by the analytical functions of van Genuchten (1980) and Mualem (1976). Then, a steady-state situation is simulated with a constant, low, and uniform infiltration rate of 0.5 mm d-1 at the soil surface, which corresponds to a similar drain discharge rate. Groundwater flow occurs only in the saturated zone below the water table. The Ks in the subsoil is manually calibrated until the measured groundwater level (Fig. 5) is simulated. The calibration of Ks values is not very sensitive for Ks values of shallower layers because the soil hydraulic characteristics of the shallower unsaturated soil layers has only a minor effect on water flow to the drains. By increasing the infiltration rate, the water table rises and a shallower soil layer becomes also (partly) saturated and the corresponding Ks value is calibrated. Steady infiltration rates of 0.5, 1, 2, 4, 6, 10, 15, 20, and 25 mm d-1 are used to find the values of Ks for the different layers. The Ks of the drain trench is fixed at and a possible entrance resistance of the drain is ignored. The calibrated Ks for each layer has to be within the measured Ks range (Fig. 4). A decrease in Ks with depth (for depth > 50 cm) is also used as a conditioning criterion because this trend is found in the measured Ks data (Fig. 4). No quantitative measure for the goodness of the calibration was used. The final results presented in Fig. 5 and Table 1 are considered to be adequate, given the conditioning criteria and the variations in the field data.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 5 Drain discharge rates as a function of the groundwater level at -6-m lateral distance from the tile-drain for the period from 17 Dec. 1991 to 8 Apr. 1992 . The measured data (points) and the steady-state simulations of the calibrated SWMS_2D model are indicated

 
de Vos et al. (1999) show that the use of the field water retention curve for the topsoil (0–25 cm) (Table 1), together with the hourly precipitation and daily evaporation data results in simulated groundwater levels and drain discharge rates in close agreement with the measurements (0 < t < 40 d).

The dynamics of the NO3 concentration in the drainage water appears to be dominated by convective transport, which determines the peaks in NO3 concentrations. A longitudinal dispersivity and a transverse dispersivity are used (de Vos et al., 1999). The order of magnitude of these dispersivities corresponds with the dispersivities of 5 to 20 cm given by van Hoorn (1981) for field-scale solute transport.


   Results
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
Discussion
 Conclusions
 REFERENCES
 
Field Measurements
Nitrate Concentration in the Drainage Water
Precipitation had a strong effect on the groundwater level (Fig. 6a and 6b) . Large fluctuations in groundwater level resulted in corresponding variations in discharge rates (Fig. 6c). The fluctuations of the NO3 concentration in the drainage water (Fig. 6d) were strongly correlated to the groundwater levels and drain discharge rates. The ammonium (NH4) concentration was measured in some drainage water samples. The NH4 concentration was nearly constant at 0.2 mg L-1, in agreement with the low NH4 content in the soil profile. So, the contribution of NH4 to the N balance was neglected.



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 6 Precipitation rate (a), groundwater level at -6 m lateral distance from the tile-drain (b), drain discharge rate (c), and nitrate concentration in the drainage water (d) during the leaching period from 17 Dec. 1991 to 8 Apr. 1992 . Measurements and simulation results are indicated

 
Nitrate Concentration Profiles
Nitrate concentration profiles measured on 17 Dec. 1991 show higher NO3 concentrations for the block at a distance of 0 to 1 m from the drain than for the other blocks (Fig. 7a) . The highest NO3 concentrations were found in the 0- to 50-cm soil depth interval. For depths larger than 100 cm, the NO3 concentrations were low. The 0- to 150-cm depth soil profile shows a statistically significant different total N content for the block at a distance of 0 to 1 m from the drain of 93 kg ha-1, compared to the average of 64 kg ha-1 for the other blocks. This difference could be caused by higher mineralization and/ or smaller denitrification rates at the drier locations closer to the drain. Differences in N uptake by the preceding sugar beet crop could be an alternative explanation. The measured NO3 concentration distribution on 17 Dec. 1991 is used as initial condition in the SWMS_2D model.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 7 Nitrate concentrations in the soil water for the blocks at different distances from the tile-drain, measured by soil sampling on 17 Dec. 1991 (t = 0; a) and 8 Apr. 1992 (t = 114 d; b), assuming all nitrate was dissolved in the soil water. On 8 Apr. 1992 the data of the topsoil are subject to errors due to the interference of fertilizer granules in the samples (see text)

 
The mineral N content on 8 Apr. 1992 (Fig. 7b) was measured just after the application of N fertilizer, which resulted in very high NO3 concentrations in the topsoil. The NO3 concentrations before fertilization for the 0- to 30-cm depth layer cannot be reconstructed accurately, due to the large variation in N content in the soil samples. At depths larger than 30 cm, no substantial differences were found in total N content as a function of the distance from the drain. The total N content in the 0- to 150-cm depth soil profile was 97 kg N ha-1, after subtracting the N fertilizer application, which is higher than on 17 Dec. 1991.

Simulations of Two-Dimensional Water Flow and Nitrate Transport
Groundwater levels are generally predicted deeper than measured, except for situations with high precipitation rates, that is, around (Fig. 6a). Simulated and measured drain discharge rates in the course of time correspond well (Fig. 6b). Predicted peak discharge rates are sometimes higher than measured discharge rates. However, in dry periods, that is, 40 < t < 80 d, the simulated drain discharge starts later than the measured discharge, which is related to the deeper simulated groundwater level and the corresponding larger water storage capacity.

Nitrate concentrations in the drainage water are overestimated for 0 < t < 90 d, and are close to the measured data for t > 90 d (Fig. 6c). The changes in NO3 concentrations in the drainage water in the course of time are simulated well. Figure 8 presents the simulated NO3 concentration distribution at three characteristic times, showing the vertical and horizontal displacement and dilution of NO3. At high infiltration rates the NO3 just above the drain is transported downwards rapidly.



View larger version (51K):
[in this window]
[in a new window]
  		Fig. 8 Simulated nitrate (NO3) concentration distribution in the flow domain for three different days during the leaching period from 17 Dec. 1991 to 8 Apr. 1992 (0 < t < 114 d)

 
Measured and simulated cumulative drain discharges for the leaching period (0 < t < 114 d) were 120 and 138 mm, and cumulative N discharges were 11 and 15 kg N ha-1, respectively.


   Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
 Discussion
Conclusions
 REFERENCES
 
Two-Dimensional Water Flow and Nitrate Transport
Peaks in NO3 concentrations correspond to the peaks in groundwater levels and drain discharge rates (Fig. 6). The temporal behavior of water flow explains the fluctuations in NO3 concentrations. Under discharge conditions with the water table close to the drain depth, like in the period 30 < t < 90 d, the origin of the discharged water was mainly from depths below the drain. The NO3 concentrations in the soil water were relatively low in that zone (Fig. 7). When the water table rose in a short period to a shallow depth (0–50 cm), like in the periods 20 < t < 25 d and 90 < t < 100 d, a substantial part of the discharged water originated from the zone where NO3 concentrations were higher. In view of the initial NO3 contents presented in Fig. 7 and possible N mineralization mainly occurring in the 0 to 25 cm of the topsoil, high NO3 concentrations in the drainage water, close to 70 mg L-1, can only originate from depths <50 cm. The discharged water is always a mixture of water from different origins, but the sharp changes in the measured NO3 concentrations indicate an abrupt shift of the stream tubes and a contribution to drainage and leaching from correspondingly different zones in the soil. Field data of chloride concentrations in the soil and the drainage water (data not shown) are opposite compared to NO3 (de Vos, 1997). High Cl concentrations are found at larger depths due to upward diffusion from the saline subsoil. At peak discharge rates Cl concentrations in the drainage water are low and at small discharge rates Cl concentrations are high. This opposite behavior compared to NO3, strengthens the interpretation of the NO3 transport processes. These interpretations are based on the field data only. A two-dimensional modeling concept is needed to simulate this temporal behavior of water flow and NO3 transport.

A field water balance is calculated for a volume of soil to a depth of 120 cm to check whether downward seepage occurred during the winter leaching period. Precipitation was and drain discharge was . The potential Makkink evapotranspiration, as given by the Royal Netherlands Meteorological Institute (Makkink, 1957; De Bruin, 1987) was used to estimate cumulative soil evaporation of . The water content in the soil profile increased by an amount of . So, downward seepage was . The measured hydraulic head gradient also indicated downward seepage. However, the groundwater level never fell far below the drain depth during relatively dry periods (Fig. 6b). All these data consistently indicate a very small downward seepage, which corresponds to the small Ks in the subsoil. The assumption of an impermeable layer as a bottom boundary condition in SWMS_2D is justified, because the 4-mm downward seepage can be ignored relative to the other components in the water balance. This hydrological situation is ideal for monitoring NO3 leaching in the drainage water, because a spatially integrated value for a catchment is found, which circumvents the problem of spatial variability encountered by other techniques to quantify NO3 leaching from a field (Addiscott et al., 1991).

The water flow part of SWMS_2D was calibrated using field data, which implies that the water balance simulations are not independent of the calibration. However, no special calibrations were used for NO3 transport. The satisfactory simulations of the NO3 fluctuations in the drainage water are an independent indication that the water flow part of SWMS_2D is describing the convective part of transport well. The large values of Ks are realistic for the Lovinkhoeve soil. The origin of the large Ks values is related to the occurrence of macropores, worm holes, old root channels, and small cracks. Evidence for the functioning of these larger pores was found during the Ks measurements and a later (1994) bromide tracer experiment (de Vos, 1997). Sharp peaks in NO3 concentrations in drainage water are often explained by preferential flow mechanisms (Nieber and Misra, 1993). However, this study shows that a two-dimensional convection–dispersion approach can sometimes also explain peaks in NO3 concentration.

Nitrogen Transformations and Nitrate Leaching
The N transformations for the 4-yr crop period 1988 to 1991 at an integrated field at the Lovinkhoeve are given in Fig. 9 (van Faassen and Lebbink, 1994). Nitrogen contents in the large organic matter pools are assumed to be constant. The yearly net N mineralization (mineralization–immobilization) is 120 kg ha-1 yr-1 and N losses to the environment are estimated at 45 kg ha-1 yr-1. Incubation experiments (Bloem et al., 1994) showed net N mineralization rates of 100 kg ha-1 yr-1 during the growing season (April–October) in the 0- to 25-cm topsoil. An average net N mineralization of 20 kg ha-1 yr-1 can thus be expected during the winter period (November–March). Corré (1995) calculated a denitrification rate of 20 kg N ha-1 yr-1 for a silt loam soil comparable to the Lovinkhoeve. So, an average N leaching of 25 kg ha-1 yr-1 is estimated from the total N losses minus denitrification, ignoring N volatilization. However, this N leaching is a rough estimate because it contains the cumulative errors of the different components of the N balance.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 9 The N balance (kg N ha-1 yr-1) in the 0–60 cm soil layer, based on a 4-yr crop rotation (1988–1991) of an integrated arable farming system (after Van Faassen and Lebbink, 1994). The young and old organic matter pools are given for the 0–25 cm soil layer

 
Different components of the N balance can also be obtained from the data of our field experiment. The N content in the 0- to 150-cm soil layer averaged for the entire field increased by 28 kg N ha-1, from 69 kg N ha-1 on 18 Dec. 1991 to 97 kg N ha-1 on 14 Apr. 1992. This increase is the net result of mineralization, immobilization, deposition, volatilization, denitrification, and leaching. The N leaching of 11 kg N ha-1 was measured, which implies a net increase of 39 kg N ha-1 in the soil profile. We assume that all N transformations occurred in the 0- to 25-cm topsoil which has a high organic matter content (de Ruiter et al., 1994).

The effect of the N-production term in the topsoil had only a minor effect on N leaching, although the N content in the soil profile increased. A simulation without this production term resulted in 2 kg N ha-1 less N leaching. However, the precipitation excess in the winter leaching period 1991–1992 was 143, which is 100 mm less than in an average winter period. Extra precipitation excess would have probably resulted in larger N leaching, depending on the actual distribution of the precipitation in time. This expectation corresponds with the estimated average N leaching of 25 kg N ha-1 yr-1. In future studies it would be worthwhile to devote more resources to frequent sampling of NO3 in the soil profile. This would allow one to capitalize on the expectation that during major leaching events of a few days, N transformations are of relative minor importance.

The SWMS_2D model can simulate the effects of different meteorological conditions, downward seepage as bottom boundary condition and a smaller drain spacing on N leaching. Absolute amounts of N leaching cannot be simulated accurately yet, because only a constant N-production is considered. A better description of N transformations can be incorporated in the model. However, with the SWMS_2D transport model as it is, relative effects of different scenarios to formulate measures to reduce N leaching can be simulated. This study again shows, like the studies of De Willigen (1991) and Feyen et al. (1998), that a mechanistic model helps to understand the field-scale processes, but that the calibration of many parameters is a problem when absolute quantitative results have to be obtained for a field situation.


   Conclusions
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
 Discussion
 Conclusions
REFERENCES
 
The temporal behavior of NO3 leaching to a subsurface tile drain in a silt loam soil is dominated by two-dimensional water flow and NO3 transport processes. Corresponding peaks in groundwater levels, drain discharge rates, and NO3 concentrations in the drainage water, can only be described well by considering the vertical distribution of the soil hydraulic characteristics and the NO3 distribution in the two-dimensional soil profile. The SWMS_2D model, calibrated for water flow and with taking into account N production in the topsoil, simulated N leaching well for the winter leaching period 1991 through 1992. For this relatively dry period N leaching of 15 kg N ha-1 was simulated vs. 11 kg N ha-1 measured. Under wetter, average meteorological conditions 25 kg N ha-1 yr-1 N leaching is expected, based on the N balance over a 4-yr period. In a silt loam soil with an organic matter content of 2.2%, net N mineralization in a winter period is 20 to 40 kg N ha-1, which is of the same order of magnitude as the average N leaching. Nitrogen transformations cannot be ignored, especially N mineralization and denitrification have to be taken into account, to describe the fate of N in the soil profile more precisely. The two-dimensional transport model SWMS_2D is a good tool to evaluate the relative effects of measures to reduce N leaching.1974; 5789


   ACKNOWLEDGMENTS
 
The authors thank J.S. Zwiers, E. Hummelink, and K. Boersma for the collection and processing of data.

As part of the Dutch Programme on Soil Ecology of Arable Farming Systems the work in the first research period, 1986 to 1990, was supported by the Netherlands Integrated Soil Research Programme (PCBB).

The later research, in the period 1994 to 1997, was part of the DGXII Environment Research Programme of the European Community, contract EV5V-CT94-0493, project PL93-1923: "Critical evaluation of selected models describing nitrate leaching and biological transformations. Quantification of the effect of various agricultural practices on nitrate pollution."

Received for publication March 15, 1999.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results
 Discussion
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
J. Simunek, M. Th. van Genuchten, and M. Sejna
Development and Applications of the HYDRUS and STANMOD Software Packages and Related Codes
Vadose Zone J., May 27, 2008; 7(2): 587 - 600.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
A. Boivin, J. Simunek, M. Schiavon, and M. Th. van Genuchten
Comparison of Pesticide Transport Processes in Three Tile-Drained Field Soils Using HYDRUS-2D
Vadose Zone J., June 21, 2006; 5(3): 838 - 849.
[Abstract] [Full Text] [PDF]

Home page
 Vadose Zone JHome page
Y. S. Onsoy, T. Harter, T. R. Ginn, and W. R. Horwath
Spatial Variability and Transport of Nitrate in a Deep Alluvial Vadose Zone
Vadose Zone J., February 1, 2005; 4(1): 41 - 54.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
W. D. Shuster, M. J. Shipitalo, S. Subler, S. Aref, and E. L. McCoy
Earthworm Additions Affect Leachate Production and Nitrogen Losses in Typical Midwestern Agroecosystems
J. Environ. Qual., November 1, 2003; 32(6): 2132 - 2139.
[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 ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (15)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by de Vos, J.A.
Right arrow Articles by Raats, P.A.C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by de Vos, J.A.
Right arrow Articles by Raats, P.A.C.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by de Vos, J.A.
Right arrow Articles by Raats, P.A.C.

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
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome