| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Biol. System Eng., Univ. of Wisconsin, Madison, WI 53706-1299 USA
b Dep. of Agric. and Biological Eng., Cornell Univ., Ithaca, NY 14850 USA
c Dep. of Agronomy, Purdue Univ., West Lafayette, IN USA
d Weed Science Lab., USDA-ARS, BARC-W, 10300 Baltimore Blvd., Beltsville, MD 20705-2350 USA
kung{at}calshp.cals.wisc.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Abbreviations: BTC, breakthrough curves • WT, water tracer
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
Among the three types of preferential flow paths, the in-situ scales, scope, and formation–destruction mechanisms of macropore flow paths are not yet clearly comprehended. Therefore, there is no evidence that impact of macropore flow on contaminant transport under field conditions can be accurately replicated and examined in laboratory studies. Consequently, the impact of macropore flow on contaminant leaching needs to be studied in field-scale experiments. The conventional sampling protocols (e.g., soil cores and solution lysimeters) used in field experiments to examine contaminant leaching were developed more than 50 years ago when the existence, mechanisms, and impacts of preferential flow were not understood. These sampling protocols implicitly assume that water-borne contaminants move through the entire soil profile. As a result, samples are collected at random locations and measured results are averaged. A sample collected near a preferential pathway will recover much more mass than those located away from a pathway. This partly explains why there was substantial scatter in sampling results that were based on the conventional sampling protocols (Ghodrati and Jury, 1992; Ju et al., 1997). When contaminants are transported through certain complex yet fixed pathways, sample locations rather than the number of samples determine the representativeness of the samples (Kung, 1990). The random samples collected by the coring method and the lysimeter method may significantly underestimate solute breakthrough in soil with preferential flow paths. In order to accurately determine the impact of preferential flow on contaminant transport, one must first detect preferential pathways and then sample along the preferential pathways (Kung, 1990; Ju et al., 1997). However, there are no tools that can nondestructively and three-dimensionally map and visualize these complex macropore flow paths. Therefore, it is necessary to use an alternative sampling protocol to assess the total leaching.
Some agricultural fields are tile-drained to control the water table. If the tiles were installed with uniform spacing and depth, and border tiles were installed to partition the input area, a field above a tile will behave like a huge, enclosed and undisturbed lysimeter. As long as the watertable is above the tile line and vertical recharge into regional groundwater is negligible, contaminants leached down from a plot above a tile can be accurately and holistically assessed from tile flow, regardless of how rapid and localized the contaminants are being transported through preferential flow paths. Richards and Steenhuis (1988), Everts et al. (1989), and Kladivko et al. (1991) showed that their tile drain monitoring facilities can be used as an alternative sampling protocol to assess the impact of preferential flow on contaminant transport.
There are a number of factors that influence pesticide field behavior: soil hydraulic properties, soil clay and organic matter contents, hydro-geologic setting, and pesticide properties being the most critical (Cohen et al., 1986). Although direct comparisons of the relative importance of these factors were essentially non-existent, it has been shown that, within uniform soils without preferential flow paths, for a pesticide with low persistence, retardation had the greatest impact on its environmental fate (Gusfafson, 1989). In field with preferential flow paths, Kladivko et al. (1991) co-applied three pesticides with different adsorbing properties and found that, after only 1 cm of net infiltration from natural precipitation occurred shortly after application, all pesticides were simultaneously detected in water samples collected from tile drains buried at 75 cm during the first day after the precipitation. One pore volume of the soil profile is 
30 cm of water. This suggested that, shortly after application, pesticides with different adsorbing properties were equally susceptible to deep leaching through preferential flow paths.
Pesticide analysis is expensive and the total number of samples that can be collected and processed are often limited by the research budget. For this reason, non-adsorbing tracers such as bromide (Br-) and chloride (Cl-) and adsorbing tracers such as rhodamine WT dye and FD&C blue no. 1 dye have been frequently used among the published studies. Conventionally, it has been assumed that: (i) the breakthrough patterns of non-adsorbing tracers such as Br- and Cl- only represent the worst-case scenario of nitrate transport through preferential flow paths; and (ii) breakthrough patterns of adsorbing tracers such as dyes can simulate those of pesticides with retardation coefficients higher and solubility lower than those of Br- and Cl-. Everts et al. (1989) used adsorbing tracer rhodamine WT dye (with Koc ranging from 1150–1650 mL g-1) and non-adsorbing tracer Br- to examine contaminant transport in simulated precipitations. By frequently sampling tile drain buried at 110 cm, they observed that both tracers had fast breakthrough patterns within the first four hour after the initiation of the irrigation. They also found that >80% of mass of the adsorbing tracer that leached out did so during the precipitation event. This confirmed that chemicals with different adsorbing properties were equally susceptible to deep leaching through preferential flow paths. The results also suggested that in order to accurately monitor the total mass of a contaminant leached out through preferential flow paths, it was critical to frequently sample during and shortly after a precipitation.
The initial soil water content can influence whether a chemical will be subject to deep leaching through preferential flow paths. The initial soil water content of Kladivko et al. (1991) was not determined, while that of Everts et al. (1989) was 0.34 cm3 cm-3, which was the maximum water holding capacity. The objective of this study was to use a tile drain monitoring facility to compare the impact of preferential flow on the breakthrough time and pattern of both adsorbing (rhodamine WT) and non-adsorbing (Br- and Cl-) tracers uniformly applied to a soil surface for two different tillage practices under both initially dry and wet conditions.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
90-cm depth along the center of each plot over 15 years ago. The slope at the soil surface varies from 3 to 5%. A monitoring facility was installed in a manhole at the lower end of each plot to continuously measure the tile flow and periodically collect water samples. Plot 3 was under no-till continuous corn, while Plot 4 was under conventional till corn–soybean rotation. The Farm is along the west shore of Lake Champlain and the soil of the site is Rhinebeck variant loam (fine, illitic, mesic Aeric Epiaqualfs) originated from glacial outwash, which consists of five distinct layers: fine sandy loam (0–21 cm); clayey loam (21–42 cm); clayey hard pan (42–64 cm); gravely loam (64–99 cm); and gravely sand (>99 cm).
In order to use tile as an alternative sampling protocol, the watertable must be above the tile line. From previous records, the flow rate of tile drainage would decrease slowly from 100 to 10 mL s-1 after a rainfall and then quickly drop from 10 to 0 mL s-1. Based on this information, the current experiments were conducted when the flow rate reached 5 mL s-1 and there was no precipitation in the 5-day weather forecast. A simulated rainfall with 7.5 mm h-1 intensity was applied to Plot 3 for 7.5 h by a solid set sprinkler system to simulate a spring shower. The irrigated area was 18 m by 65 m, which covered the entire width of the plot and was 5 m from the lower end of the plot. The riser height of the sprinkler head was 1.5 m and the height of corn was 1.0 m (there was an unprecedented amount of natural precipitation in early summer that postponed the experiments well past crop emergence.)
Shortly after the irrigation started, non-adsorbing tracer KBr (15.3 kg Br-) and adsorbing tracer rhodamine WT (0.65 L solution with 15% by weight) dissolved in 150 L water were injected into the irrigation water at the pump, which was located 1.6 km (about a mile) from the plots. By visually observing the color of the irrigation water at the sprinkler nozzles, it was found that the first pulse of tracers arrived at the plot between 72 min and 96 min after the irrigation started. About 4 h after injecting the first pulse of tracers, KCl (19.6 kg Cl-) and rhodamine WT (0.65 L solution with 15% by weight) were again dissolved in 150 L water and injected into the irrigation water at the pump. This pulse of tracers arrived at the plot between 311 min and 335 min after the irrigation started.
The first pulse was to test the impact of preferential flow on the transport of non-adsorbing and adsorbing tracers under relatively drier soil conditions, while the second pulse was to test under wetter conditions. Identical experiments were also conducted in Plot 4, except the irrigation intensity was reduced to 5 mm h-1 for 7.5 h to reduce surface runoff. There was no facility to measure the water and chemicals lost in surface runoff. Furthermore, fewer tracers were applied in Plot 4; i.e., 8.27 kg Br- and 0.35 L rhodamine WT with 15% by weight were applied in the first pulse, and 12.7 kg Cl- and 0.35 L rhodamine WT with 15% by weight were applied in the second pulse. The timing and method of the tracer application in Plot 4 were identical to those in Plot 3.
During the first 12 h after the irrigation started, water samples were collected from the tile every 6 min. From 12 to 18 h, water samples were collected every 15 min. From 18 to 24 h, water samples were collected every 30 min. During the second and the third days, water samples were collected every 2 h. The samples were stored at approximately 5°C until analyzed, except during a short period of shipping. Flow rate of the tile drain was continuously measured during the 3-d period by the tipping-bucket method. The Br-, Cl-, and nitrate in the water samples were analyzed using the method discussed in Kung et al. (2000), while rhodamine WT was analyzed by using fluorolite fluorescent meter with 425 nm excitation and 515 nm reflection filters. The detection limits of Br-, Cl-, and nitrate were 0.25, 0.5, and 0.5 mg L-1, respectively, while that of the rhodamine WT was 1 µg L-1.
|
Results and discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
|
2 d for the applied tracers to arrive at the tile line under a free drainage condition with unit gradient (i.e., v = q/
= 7.5 mm h-1/0.4 = 18.9 mm h-1 and t = L/v = 0.9 m/18.9 mm h-1 2 d). Because samples were taken once every 6 min, whether the adsorbing tracer rhodamine WT had the exact same arrival time as that of the non-adsorbing tracer Br- was unknown (i.e., it was possible that one tracer arrived earlier than the other between adjacent sampling interval). Nevertheless, it was not expected that a highly adsorbing tracer like rhodamine WT arrived at a tile buried at 0.9-m depth at only 24 min after its application. In short, the rapid arrival time unambiguously demonstrated that preferential flow dictates the initial transport of an adsorbing contaminant. It has been conventionally conceptualized that water enters into preferential flow paths such as macropores only when the soil surface is very wet (i.e., soil surface was nearly saturated with localized ponding). The gravimetric water content of the top 30 cm of soil, based on 9 cores taken from an adjacent area before the irrigation started, was between 11 and 15%, while the saturated water content of the soil was 36%. About 15 mm irrigation water was applied during the first 2 h and this was far from enough to saturate the top 30 cm soil. According to our visual observation, there was neither localized ponding nor surface runoff during the first two hours after irrigation started. The first pulse of tracers arrived at the plot between 72 min and 96 min after the irrigation started. This indicated that contaminant transport through preferential flow paths occurred when the top 30 cm soil was under unsaturated conditions.
Figure 1 also shows that the BTC of both Br- and rhodamine WT reach a peak at the identical time (i.e., 126 min after the irrigation started or 54 min after tracer application). From the arrival to the peak of the tracers applied in the first pulse, tile flow increased very quickly and the concentrations of both Cl- and nitrate continuously decreased. This suggested that more and more preferential flow paths became hydraulically active. As an adsorptive tracer transported down through a preferential flow path, it was expected that some would be retarded along the wall of the path. For this reason, it has been conventionally assumed that because Br- have essentially no retardation effect, they represent the worst-case scenario of contaminant transport through preferential flow paths. However, the similarity of the initial BTC of the Br- and rhodamine WT suggested that an adsorbing tracer was as susceptible to fast and deep leaching through preferential flow paths as the non-adsorbing tracers. This indicated that the chemical properties of the tracer had less effect on the initial transport through preferential flow paths.
As shown in Fig. 1, the Cl- and rhodamine WT applied in the second pulse again arrived at the tile drain at the same time (i.e., 324 min after irrigation began or 13 min after tracer application). The BTC of these two chemicals had similar shape and peaked at the same time (354 min after irrigation began or 43 min after application). This fast breakthrough confirmed that it was the water dynamics, instead of the chemical properties of the tracer, that dictated the initial transport of contaminants through preferential flow paths. The rate of tile flow at the time of the Cl- arrival was about 10 times higher than that at Br- arrival. This suggested that many more preferential flow paths had became hydraulically active. The fact that the Cl- arrival time was much shorter than that of Br- suggested that the newly hydraulically active preferential flow paths had either larger pores or straighter paths. When a soil profile becomes wet, because the matric potential gradient decreases, it is logical that preferential flow paths with larger pores would become hydraulically active. There is no reason that a straighter preferential flow path would became hydraulically active later than a more tortuous path with identical equivalent pore radius. Therefore, the shorter arrival time was probably because some Cl- was transported through preferential flow paths with larger pores, which were not hydraulically active earlier.
Figure 1 also shows that the BTC of the Br- and rhodamine WT have similar patterns during the first 36 min after initial arrival at the tile drain, while those of Cl- and rhodamine WT have almost the identical shape during the first 140 min after initial arrival. This indicated that, the wetter the soil profile, the more tracers had entered into preferential flow paths with larger pores and, hence, the longer the transport of an adsorptive tracer will behave like that of a non-adsorptive tracer. This confirmed that (i) the preferential flow paths had a spectrum of pore sizes and that (ii) some preferential flow paths with larger pores were initially not hydraulically active when their soil profile was dry.
When comparing the tails of the BTC of the adsorbing and non-adsorbing tracers applied in both pulses, it was clear that the adsorbing tracer rhodamine WT trailed off much more quickly. The tracers arriving at the tile early must be transported through the preferential paths with the larger pores. An adsorbing tracer that entered into preferential flow paths with larger pores would have less residence time in the unsaturated zone and a higher probability to reach the watertable without being adsorbed. On the other hand, tracers arriving at the tile at the tail of the BTC were transported either through the preferential paths with the smaller and/or more tortuous preferential pores or through soil matrix pores. Within the smaller and/or more tortuous pores, an adsorbing contaminant would stay in soil longer and have a higher probability of being retarded. As a result, the chemical adsorption would play a more important role in determining the total deep leaching when a contaminant is transported through smaller and/or more tortuous pores. Under an extreme case, where soil is homogenized and repacked in a laboratory column to destroy preferential flow paths, the fast breakthrough would be eliminated and the chemical properties of the contaminant would completely dominate the transport. These results were consistent with the fate of five pesticides (with different retardation factors) in field experiments by Gish et al. (1990).
The tracer concentrations and tile hydrograph from Plot 4 are shown in Fig. 2 . Tile flow started to quickly increase at 140 min after irrigation. The background concentrations of the Br- and rhodamine WT from the tile flow were below their detection limits, while those of the Cl- and nitrate were 36 and 75 mg L-1, respectively. Again, the concentrations of the Cl- and nitrate decreased initially by dilution, as those shown in Plot 3. The later water arrival in Plot 4 was mainly because the irrigation intensity in Plot 3 was 50% higher (5 mm h-1 in Plot 4 and 7.5 mm h-1 in Plot 3). Bromide and rhodamine WT were detected from the water sample collected at 162 and 168 min after the simulated rainfall started (or 90 and 96 min after the application of the first pulse of tracers), respectively. These arrival times were slower as compared to that in Plot 3 (24 min after tracer application). Because the irrigation intensity in Plot 3 was 50% higher, the surface soil in Plot 3 was wetter when the first pulse of tracers was applied, and hence the tracers in Plot 3 had a greater chance to enter and be transported quickly through preferential flow paths.
|
Because less water and tracers were applied to Plot 4, the peak of the tile hydrograph, as well as the peak concentrations of the tracers shown in Fig. 2, are lower than those in Fig. 1. In Fig. 1, the two peaks of the rhodamine WT have almost identical concentration. In Fig. 2, however, the first peak of the rhodamine WT has a much lower concentration than that of the second peak. Within each plot, the total amount of rhodamine WT applied in the two pulses was identical. Among the four pulses of tracer applications in two plots, the least amount of irrigation water (6 mm) was applied when the tracers from the first pulse in Plot 4 were applied. As a result, the adsorbing rhodamine WT from the first pulse applied to Plot 4 had a higher probability of entering smaller pores and being retarded.
Figures 1 and 2 show that as the tile flow starts to decrease, the concentrations of the rhodamine WT decrease quickly by two orders of magnitude. From tracer concentration and tile flow, mass flux of the leached tracer can be calculated. The result showed that, of the total mass leached out from the tile, >90% of rhodamine WT was leached during the first 20 h. This suggests that the vast majority of the deep leaching of absorptive pesticides occurs during the first day after a precipitation event. Currently, there is no guideline on when and how frequently to collect samples in the coring and lysimeter sampling protocols. Results from this study confirmed the observation by Everts et al., (1989) that, in order to accurately assess the deep leaching of adsorbing pesticides, one must intensively sample during the first day after a major precipitation event.
The mass recoveries of Br- from tile flow in Plots 3 and 4 were 3.41 and 3.51% of the total mass injected into the irrigation water at the irrigation pump, respectively; while those of rhodamine WT were 0.79 and 0.194%, respectively. Note that the real mass recoveries were unknown because: (i) the pipeline from the lake to the site was 1.6 km long and there were several leakage spots along the joints and vents; (ii) some tracers left the plots through surface runoff, which occurred at about 4 and 5.4 h after irrigation started in Plots 3 and 4, respectively; and (iii) the corn canopy was 1 m in height and unavoidably intercepted some of the tracers applied. Therefore, although the exact amount of each tracer injected into the irrigation water was known, the exact amount of mass actually infiltrated into the soil surface could not be accurately estimated. Furthermore, some tracers that reached the water table probably had flowed within the saturated zone and drained toward the adjacent bordering tiles, which were not monitored. Nevertheless, it was clear that much more nonadsorbing tracer leached out than the adsorbing tracer. This was mainly because the tail of the adsorbing tracer trailed off much more quickly. The total mass recoveries of Cl- were not calculated because the soil profile had high initial Cl- concentration. It was impossible to partition the newly applied Cl- from the old Cl- in the soil.
|
Conclusions |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication February 2, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResults and discussionConclusionsREFERENCES |
|---|
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]
|
|
|
|
 |
|
 G. Kahl, J. Ingwersen, P. Nutniyom, S. Totrakool, K. Pansombat, P. Thavornyutikarn, and T. Streck Micro-Trench Experiments on Interflow and Lateral Pesticide Transport in a Sloped Soil in Northern Thailand J. Environ. Qual., June 27, 2007; 36(4): 1205 - 1216. [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]
|
|
|
|
 |
|
 L. Alletto, Y. Coquet, P. Vachier, and C. Labat Hydraulic Conductivity, Immobile Water Content, and Exchange Coefficient in Three Soil Profiles Soil Sci. Soc. Am. J., June 21, 2006; 70(4): 1272 - 1280. [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]
|
|
|
|
|
|
Y.-J. Kim, C. J. G. Darnault, N. O. Bailey, J.-Y. Parlange, and T. S. Steenhuis Equation for Describing Solute Transport in Field Soils with Preferential Flow Paths Soil Sci. Soc. Am. J., March 1, 2005; 69(2): 291 - 300. [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]
|
|
|
|
|
|
T. J. Gish, K.-J. S. Kung, D. C. Perry, J. Posner, G. Bubenzer, C. S. Helling, E. J. Kladivko, and T. S. Steenhuis Impact of Preferential Flow at Varying Irrigation Rates by Quantifying Mass Fluxes J. Environ. Qual., May 1, 2004; 33(3): 1033 - 1040. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. J. Gish, W. P. Dulaney, K.-J. S. Kung, C. S. T. Daughtry, J. A. Doolittle, and P. T. Miller Evaluating Use of Ground-Penetrating Radar for Identifying Subsurface Flow Pathways Soil Sci. Soc. Am. J., September 1, 2002; 66(5): 1620 - 1629. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Vanderborght, P. Gahwiller, and H. Fluhler Identification of Transport Processes in Soil Cores Using Fluorescent Tracers Soil Sci. Soc. Am. J., May 1, 2002; 66(3): 774 - 787. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. B. Jaynes, S. I. Ahmed, K.-J. S. Kung, and R. S. Kanwar Temporal Dynamics of Preferential Flow to a Subsurface Drain Soil Sci. Soc. Am. J., September 1, 2001; 65(5): 1368 - 1376. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.-J.S. Kung, E.J. Kladivko, T.J. Gish, T.S. Steenhuis, G. Bubenzer, and C.S. Helling Quantifying Preferential Flow by Breakthrough of Sequentially Applied Tracers: Silt Loam Soil Soil Sci. Soc. Am. J., July 1, 2000; 64(4): 1296 - 1304. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
