| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS-National Soil Tilth Lab., 2150 Pammel Dr., Ames, IA 50011-3120
b Agricultural & Biosystems Engineering, Iowa State Univ., Ames, IA 50011
c Dep. Soil Science, Univ. of Wisconsin, Madison, WI 53706
* Corresponding author (jaynes{at}nstl.gov)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Abbreviations: DOY, day of year • PF, pentafluorobenzoate • TF, o-tri-fluoromethylbenzoate • DF, difluorobenzoate
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Little is known about the stability of preferential flow pathways over time. Using large undisturbed soil blocks, Ogden et al. (1999) showed that spatial outflow patterns changed between irrigation events and were more variable in no-till soils where macropores are more likely to be preserved than in plow-till soils (Andreini and Steenhuis, 1990; Isensee et al., 1990). In a recent study, Lennartz et al. (1999) showed that preferential flow accounted for much of the Br mass loss to a tile at a field site over 3 yr. However, while the Br leaching pattern was consistent for the first 2 yr, a greater contribution from matrix flow was observed for the third year. They attributed this change in leaching pattern to differences in precipitation patterns shortly after Br application, with little rainfall occurring in the third year, rather than to changes in preferential pathways. The dependency of chemical transport via preferential flow on the time elapsed between chemical application and leaching has also been shown in laboratory columns (Kluitenberg and Horton, 1990; Edwards et al., 1992) and for chemigation of field soils (Jaynes et al., 1992).
Even less is known about temporal dynamics of preferential flow within a single leaching event. Are preferential flow pathways stable over time? Does macropore flow develop or change over time even if the boundary conditions remain constant? Knowing the answers to these questions would help in the development of more realistic conceptual constructs of preferential flow mechanisms and should lead to the development of more accurate simulation models of water and chemical movement where preferential flow is significant. Thus, the objectives of this study were to quantify the contribution of preferential flow to the overall chemical movement to a field drainage tile during the first few weeks after application and to investigate how the timing of chemical application relative to the start of irrigation affects the temporal behavior of preferential flow during an irrigation event.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
The experiment was conducted during the growing season of 1998 within a field plot that was under no-till continuous corn (Zea mays L.) management since 1984. In late June, a solid set sprinkler irrigation system (SM20H 17° nozzles, Rain Bird, Glendora, CA1) was installed in a regular grid pattern with a 6-m spacing covering an area 24.4-m long by 42.7-m wide, centered over a 1.2-m deep tile (Fig. 1) . Corn rows and all field traffic were perpendicular to the direction of the tile. Preliminary tests had shown that this arrangement of sprinklers provided the most uniform water application pattern with the nozzles used. To prevent surface ponding, water was pumped to the sprinklers at a controlled pressure to provide a target irrigation rate of 4 mm h-1. The volume of water pumped was periodically measured with a mechanical flow meter and recorded versus time, and total irrigation depth was confirmed by measuring depth of irrigation intercepted by catch cans placed in each quadrant of the plot. Rain fell on several days after irrigation and before the cessation of tile flow. Rainfall was measured onsite with a tipping bucket rain gauge.
|
The center of the plot was drained by a tile located 1.2-m below surface that empties into a sump at the lower end of the plot. Flow rate from the tile was measured by pumping from the sump through a FP-5300 paddle wheel flow meter (Omega, Stamford, CT) and recording pumped volume every minute with a datalogger. Tile water samples were collected for chemical analysis by pumping a 300 mL sample from the tile outlet using a peristaltic pump autosampler (model 3700 ISCO, Lincoln, NE). A 5 min interval between water samples was used for the first 10 h after irrigation started. Thereafter, composite samples were collected consisting of either two or three subsamples collected every 10 min—increasing to every 60 min, 9 d after the start of the experiment—until 14 d after the start of irrigation. All water samples were stored in the dark at 4°C until chemical analysis.
Twenty-one days after irrigation, a 1.2-m deep soil core was taken every 3 m along the tracer strip for a total of 8 cores. Two additional cores were taken outside the plot to test for background levels of the chemicals used. Soil cores to a depth of 0.15 m were also taken adjacent to the strip at distances of 0.5, 1, and 2 m to check for movement of tracers in surface runoff from the strip. Soil cores were collected by pushing a 38.1-mm diam. steel soil probe fitted with a removable acetate liner into the soil with a hydraulic ram. The soil core and liner were removed from the steel probe, capped on each end, and stored at -10°C until tracer extraction. The frozen soil cores were cut into 150-mm long sections, removed from the liners, thawed, and mixed by hand. The soil was weighed and added to an Erlenmeyer flask with an approximately equal mass of 0.001 M CaSO4 solution and shaken on a wrist shaker for 15 min. Twenty mL of solution were filtered through 0.45 µm filter paper for chemical analysis and the remainder of the sample was dried at 104°C for 48 h and weighed to determine initial water content.
Chemical Analysis
Analysis of the tracers was performed on a Dionex Series 4500i ion chromatograph (Dionex, Sunnyvale, CA) using the method described by Bowman and Gibbens (1992). For the fluorobenzoates, a SAX column (Regis Chemical Co., Morton Grove, IL) was used with a mobile phase consisting of 30 mM KH2PO4 buffer, adjusted to a pH of 2.85 with H3PO4, and 200 mL L-1 acetonitrile as an organic modifier. Flow rate was 1 mL min-1 and the detection wavelength was 205 nm. Bromide could not be quantified with the above procedure because of interferences caused by high NO3 levels. Instead, Br- was determined using a Dionex AG9 guard column (Dionex, Sunnyvale, CA) followed by an AS9 separator column (Dionex, Sunnyvale, CA). The eluting solution was 1 mM Na2CO3 and 0.75 mM NaHCO3 at a pH of 10.4 with 12.5 mM H2SO4 used for suppression. Flow rate was 1 mL min-1 and electrical conductance was measured with a conductivity detector. The quantitation limit for both Br- and benzoates was 0.1 mg L-1 in the extract solutions. Concentrations below the quantitation limit were assigned values of zero in all analyses. Tracer concentrations in soil samples were calculated by multiplying measured concentrations by the sum of the mass of soil water plus added CaSO4 solution and dividing by the mass of soil water.
The extraction procedures for herbicides from water and soil samples are described in Hatfield et al. (1999). Water samples were passed through an SPE cartridge (Bond Elute LRC, 500 mg C-18, Varian, Harbor City, CA) and eluted with 2 mL of ethyl acetate using a Zymark robotic arm (Zymark Corporation, Hopkinton, MA) and system V controller coupled to a system consisting of a custom-built sample rack. Soil samples were extracted with methanol and water followed by cleanup on a solid phase extraction cartridge with elution using ethyl acetate (Koskinen et al., 1991). A Zymark robotic system was also used for this procedure. Herbicide samples were analyzed by gas chromatography (GC)/mass spectroscopy (MS) using a SIM Hewlet Packard 5970 (Hewlett Packard, Palo Alto, CA). Quantitation limits for both atrazine and alachlor for the conditions used were 0.2 µg L-1 and values below this were assigned zero.
Model Simulation
The two-dimensional model HYDRUS–2D (
im
nek et al., 1999) was used to simulate the leaching experiment. HYDRUS–2D numerically solves Richard's equation for variably saturated soil and the convective–dispersive equation for solute transport. The modeled system was 17.4-m wide and 3.0-m deep. The 1.2-m deep tile in the middle of the plot was simulated with a boundary node surrounded by four regular square elements with hydraulic conductivities adjusted according to the electric analog approach of Vimoke et al. (1963) and Fips et al. (1986) and implemented by HYDRUS–2D. The soil was divided into four layers, 0- to 0.15-, 0.15- to 0.3-, 0.3- to 0.6-, and 0.6- to 3-m deep, and assigned hydraulic properties as described by Mohanty et al. (1994) and Mohanty et al. (1998) who measured soil hydraulic properties at this site. A longitudinal dispersivity of 15 cm from Rice et al. (1986) was used for the global dispersion process represented by the averaged soil core results. All other model parameters were estimated internally by HYDRUS–2D. Initial moisture conditions were defined by simulating the draining of the profile from an initially wet state until simulated tile discharge was within 0.1 mm h-1 of the measured tile discharge before irrigation started. Initial solute conditions consisted of a concentrated layer of solute within the top 1 cm of the soil profile within a 1-m wide zone whose center was offset laterally from the tile by 1.5-m. Input solute concentrations were adjusted to give a mass of solute equal to that applied in the field. Boundary conditions included no flow boundaries on the sides, specified irrigation or rainfall rates at the surface, and a bottom boundary with a variable seepage rate as permitted by the model. A total of 1551 nodes and 2944 quadrilateral elements were used with a finer discretization near the soil surface, in the zone of chemical placement, and surrounding the tile to improve descriptions of abrupt changes in pressure gradients and solute concentration.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
|
6 h after the start of irrigation (Fig. 3 and 4)
Discharge rates increased from 0.13 mm h-1 at the start of irrigation to over 0.8 mm h-1 immediately after the end of irrigation. Discharge rates then decreased sharply within the first hour after irrigation ceased. Increases in discharge rate around DOY 186 and 187 were in response to several large rainfall events on those days. Tile discharge ceased by DOY 195 due to lack of additional water inputs and cumulative evaporation and drainage. Total water discharged through the tile drain equaled 44.1 mm or about 40% of the total irrigation and rain. Much of the difference between applied and drained water volumes was probably because of canopy interception and evapotranspiration during the 21-d period (135 mm of pan evaporation were measured during the 21 d at a location 0.5-km distant, R. Carlson, personal communication). However, some of the difference between water applied and discharged by the tile may be attributed to seepage below or laterally away from the tile.
|
|
Upon reaching a maximum, the conservative tracer concentrations decreased after the cessation of irrigation. The concentrations increased and decreased again on DOY 186 and 187 when the tile discharge responded to subsequent rain events. By the end of the observation period, all of the conservative tracers except Br were again at concentrations below the detection limit of 0.1 mg L-1. This parallel increase and decrease of tracer concentration and tile discharge rate was identified as event-driven flow by Lennartz et al. (1999) and has been identified in other studies (Kladivko et al., 1991; Jarvis et al., 1991; Laubel et al., 1999; Kung et al., 2000a,b) as a consequence of preferential flow.
Drainage results for the first 10 h of irrigation clearly show an effect of tracer application timing on tracer arrival in the tile discharge (Fig. 4). About 102 min after the start of irrigation, Br was detected in the tile effluent. First arrival of Br occurred before any measurable increase in tile discharge rate (<0.008 mm h-1 change). Bromide concentrations increased for 3 h then abruptly decreased even as irrigation continued, before increasing again. Assuming no lateral displacement of Br at the surface because of the lack of surface ponding, the Br contained within this first peak was equivalent to about 5 mm2 of the 24.4-m2 area sprayed with tracer. Thus, this first peak and breakthrough of Br could have been because of a single macropore within the tracer application strip. The subsequent increase in Br concentration would be because of the contribution of an increasing number of other preferential flow paths.
The other three conservative tracers, PF, TF, and DF, appeared in the tile effluent in the same order they were applied. To varying degrees, these tracers exhibited the same behavior as Br during the first 10 h—first increasing, then decreasing, before increasing again during the irrigation. This consistent pattern for all tracers implies that at least the most rapid pathway for chemical movement through the soil profile was active during the entire irrigation.
As the irrigation continued, tracer transport along the preferential pathways became increasingly faster. Thus, while it took a little <102 min after irrigation started before Br appeared in the tile, it took only 33 and 35 min for PF and TF, and only 15 min for DF to appear after they were applied. Assuming vertical flow only, this gives tracer velocities of between 0.000196 and 0.00133 m s-1, which are in the range for tracer velocities in macropores found by Laubel et al. (1999). Moreover, the depth of applied water required to leach the tracers to the tile decreased with order of tracer application. While it took 7.1 mm of irrigation to leach Br to the tile, it took 2.3 and 2.4 mm for PF and TF, and only 1 mm of irrigation to leach DF to the tile. This is the same pattern observed by Kung et al., 2000b where markedly less time was required for tracers applied later in the irrigation to be detected in the tile discharge. The overall similarity of the leaching pattern for the different tracers while the rate of leaching accelerated during the experiment implies that while the preferential pathways may remain the same, water and tracer movement through the pathways increases as irrigation progresses and the soil becomes progressively wetter (Kung et al., 2000b).
First detection of alachlor in tile effluent occurred at the same time as Br, and atrazine was detected 10 min earlier (Fig. 5 and 6) . However, because the detection limit for the herbicides was 500 times lower than for Br while only about 40 times more Br mass was applied to the plot, the detection/application ratio was about 12.5 times greater for the herbicides than for Br. Adjusting for this difference in detection sensitivity to mass applied, gives the first detection of atrazine comparable with Br about 30 min after the detection of Br and the first detection of alachlor about 40 min after the detection of Br. Thus, after accounting for the differences in application mass and detection sensitivity, herbicide transport to tile outlet was retarded compared with Br, which is expected for matrix flow given the affinity of these herbicides to sorb to soil [distribuion coefficient expressed on the basis of organic C (Koc) = 91 L kg-1 for atrazine and Koc = 157 L kg-1 for alachlor, Kladivko et al., 1991] and the high organic C content of these soils (20–30 g kg-1, Cambardella et al., 1994). Apparently, the herbicides also interacted with the soil along the preferential flow paths and were thus retarded with respect to Br. In addition, atrazine concentrations were about two times higher than alachlor in the tile effluent throughout the monitoring period, even though slightly more alachlor was applied to the tracer strip than atrazine. In other field and watershed studies where similar soils occur, the same trend has been found with atrazine measured in subsurface drains at appreciable concentrations (>3 µg L-1), while alachlor is rarely detected (Jayachandran et al., 1994; Jaynes et al., 1999; Moorman et al., 1999). This difference was attributed by Kladivko et al. (1991) to the higher soil adsorption affinity for alachlor versus atrazine.
|
|
1.5 L kg-1). The overall breakthrough patterns of the herbicides in the tile effluent were very similar to Br as well—increasing during the irrigation, decreasing during recession after irrigation ended, and again increasing and decreasing during the rainfall events of DOY 186 to 187 (Fig. 5). Variations in herbicide concentrations in tile discharge during the first 10 h of irrigation also mimicked Br (Fig. 6). Herbicide concentrations first increased, then decreased before continuing to increase during the first 6 h of irrigation. This early nonmonotonic behavior common to all of the chemical tracers represents transport along common preferential flow paths.
Soil Residues
Recovered chemical mass per unit depth was computed for each of the eight soil cores and averaged for the eight soil depth increments (Fig. 7 and 8)
. For the first three conservative tracers applied, the greatest mass was recovered in the 30- to 45-cm depth increment. The last conservative tracer applied, DF, had the greatest amount of mass recovered in the 15- to 30-cm depth increment. No PF, TF, or DF was recovered in the soil samples collected outside of the tracer strip. Bromide recovered outside the tracer strip was <0.01 g g-1 of the Br recovered from the surface layer within the tracer strip. Thus, little lateral movement of tracer in runoff occurred either during the low intensity irrigation or subsequent rainfall events.
|
|
Herbicide mass recoveries demonstrated a different pattern than the conservative tracers (Fig. 8). Herbicide mass decreased exponentially with depth except for a slight increase in the 91- to 106-cm layer. This increase may have been because of a change in soil texture or bulk density, although earlier investigations at this site have not reported differences at this depth (Kanwar et al., 1989). The pattern of herbicide mass residues in the soil profile are consistent with a sorbing chemical, where most of the chemical is sorbed to the surface layers where soil organic C is highest. The herbicide residue data is in contrast to the tile effluent results where the herbicides arrived very rapidly. Apparently, most of the herbicide and conservative tracers moved within the soil matrix in a manner consistent with matrix flow in a porous medium without preferential flow paths. However, a small portion of the chemicals (<1%) moved rapidly through preferential pathways and arrived at the tile in <2 h.
Total mass recoveries of the chemical tracers used were computed from the measured soil residues and tile effluent losses. Mass recoveries in the soil were 1.01 kg kg-1 for Br, 0.90 kg kg-1 for PF, 0.83 kg kg-1 for TF, 0.84 kg kg-1 for DF, 0.64 kg kg-1 for atrazine, and 0.37 kg kg-1 for alachlor. Mass recoveries for the conservative tracers were better than observed in similar field experiments (Jaynes et al., 1992; Starr and Glotfelty, 1990). Mass recoveries for the herbicides were considerably lower than the conservative tracers because of dissipation of the herbicides during the 20 d between application and recovery. Adjusting the expected mass remaining after 20 d using a 60-d half-life for atrazine and a 15-d half-life for alachlor (Baker et al., 1992), gives mass recoveries in the soil of 0.80 kg kg-1 for atrazine and 0.94 kg kg-1 for alachlor which are comparable with the conservative tracers.
Mass recovered in tile discharge was considerably lower than in the soil, equaling 0.03 kg kg-1 for Br, 0.05 kg kg-1 for PF and TF and 0.07 kg kg-1 for DF. Less relative mass was recovered for Br perhaps because it was applied before irrigation started as noted earlier. Relatively more DF was recovered in the tile discharge, which parallels the more rapid travel time of this tracer which was applied last. Greater loss of DF in tile discharge may have been because of greater transport along preferential pathways during the latter stages of the irrigation.
Herbicide mass recoveries in the tile effluent were considerably lower than the conservative tracers. Mass recovery of atrazine in the effluent was 0.007 kg kg-1 and 0.003 kg kg-1 was recovered for alachlor. Kladivko et al. (1991) also observed more atrazine leaching to subsurface drains than alachlor and attributed it to the greater adsorption affinity of alachlor to soil. Thus, although the herbicides traveled along preferential flow pathways quickly, interaction between the herbicide and the soil did occur.
Overall mass recoveries for Br were 1.04 kg kg-1, 0.94 kg kg-1 for PF, 0.88 kg kg-1 for TF, and 0.91 kg kg-1 for DF after summing soil residues and losses in tile effluent. Adjusting for degradation of the herbicides gives overall mass recoveries of 0.81 kg kg-1 for atrazine and 0.94 kg kg-1 for alachlor. These recoveries are excellent compared with other field studies given the uncertainties in half-lives and the other possible dissipation pathways such as volatilization that were not considered. Thus, comparable mass recoveries were found for all of the chemical tracers used. However, relatively more mass was lost in tile effluent for the conservative tracers than for the herbicides. Nevertheless, during no-till corn production, atrazine leaching to tiles in this soil has been found to periodically exceed 3 µg L-1, the maximum contaminant level set for drinking water by the US EPA, with the transport process attributed to preferential flow within the soil (Jayachandran et al., 1994).
Modeling
The rapid arrival of the chemical tracers can only be explained by the presence of preferential pathways in this soil. However, the pattern of increasing and decreasing chemical concentration in tile effluent in response to the irrigation and rainfall events was somewhat surprising. This pattern may have been the result of the geometry of the experimental design—a narrow tracer strip offset from above the tile within a much larger irrigated and drained area subject to intermittent irrigation and rain. A simulation using the two-dimensional, variably saturated, convective–dispersive, numerical model HYDRUS–2D (imnek et al., 1999) was used to test whether the patterns observed were due in some way to the geometry or boundary conditions of the experimental design.
Model results for the residual mass of Br in the soil under the center of the tracer strip 20 d after tracer application are included in Fig. 7. Overall, the simulated mass profile matched the measured profile within observation error except near the surface where more Br was found in the field. The simulated water discharge from the tile reproduced the measured patterns (results not shown), but predicted 50% greater total volume discharged over the first 14 d than measured which should have contributed to greater predicted chemical leaching. However, even with more drainage, the simulated concentrations of Br in the tile effluent were much less than those measured (Fig. 3). Inability of the convective–dispersive type model to reproduce the tracer concentration pattern observed in the tile effluent while matching soil residue data well, confirms that a small portion of the tracers were leached via preferential pathways—a mechanism not included in the model—and that observed effluent concentration patterns were a consequence of preferential leaching within the soil profile.
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Herbicides applied with Br also arrived at the tile within the first 2 h of the irrigation indicating transport via preferential pathways. However, relatively less herbicide mass and comparatively later arrival times of herbicides than Br indicated that there was interaction between the herbicides and the soil along the preferential pathways. Conservative tracers applied during the latter stages of irrigation arrived at the tile faster than tracers applied either immediately before or during early stages of the irrigation. The last tracer, applied 6 h after the start of irrigation, took only 15 min and 1 mm of irrigation water to travel the 120 cm between the soil surface and the tile. Overall, transport of solutes along preferential pathways appears complicated by the existence of pathways with different solute transport velocities and by a temporal trend of increasing transport velocities as irrigation progresses. These characteristics of preferential flow need to be incorporated into conceptual and numerical models of solute transport in these soils.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
Received for publication May 17, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSNOTESRESULTS AND DISCUSSIONSUMMARY AND CONCLUSIONSREFERENCES |
|---|
imnek, J., M. ejna, and M.Th. van Genuchten. 1999. The HYDRUS-2D software package for simulating the two-dimensional movement of water, heat, and multiple solutes in variably-saturated media. U.S. Salinity Lab., USDA–ARS, Riverside, CA.This article has been cited by other articles:
|
|
 |
|
  H. H. Gerke, J. Dusek, T. Vogel, and J. M. Kohne Two-Dimensional Dual-Permeability Analyses of a Bromide Tracer Experiment on a Tile-Drained Field Vadose Zone J., August 23, 2007; 6(3): 651 - 667. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.-J.S. Kung, E. J. Kladivko, C. S. Helling, T. J. Gish, T. S. Steenhuis, and D. B. Jaynes Quantifying the Pore Size Spectrum of Macropore-Type Preferential Pathways under Transient Flow Vadose Zone J., August 24, 2006; 5(3): 978 - 989. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
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]
|
|
|
|
 |
|
 A. Gaur, R. Horton, D. B. Jaynes, and J. L. Baker Measured and Predicted Solute Transport in a Tile Drained Field Soil Sci. Soc. Am. J., April 19, 2006; 70(3): 872 - 881. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.-J. S. Kung, M. Hanke, C. S. Helling, E. J. Kladivko, T. J. Gish, T. S. Steenhuis, and D. B. Jaynes Quantifying Pore-Size Spectrum of Macropore-Type Preferential Pathways Soil Sci. Soc. Am. J., June 28, 2005; 69(4): 1196 - 1208. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Herbst, W. Fialkiewicz, T. Chen, T. Putz, D. Thiery, C. Mouvet, G. Vachaud, and H. Vereecken Intercomparison of Flow and Transport Models Applied to Vertical Drainage in Cropped Lysimeters Vadose Zone J., April 25, 2005; 4(2): 240 - 254. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Kohne and H. H. Gerke Spatial and Temporal Dynamics of Preferential Bromide Movement towards a Tile Drain Vadose Zone J., February 1, 2005; 4(1): 79 - 88. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W. Malone, M. J. Shipitalo, R. D. Wauchope, and H. Sumner Residual and Contact Herbicide Transport through Field Lysimeters via Preferential Flow J. Environ. Qual., November 1, 2004; 33(6): 2141 - 2148. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kohler, A. Kohler, K. C. Abbaspour, M. Fritsch, and R. Schulin Using Simple Bucket Models to Analyze Solute Export to Subsurface Drains by Preferential Flow Vadose Zone J., February 1, 2003; 2(1): 68 - 75. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS |
