JEQ Grow Your Career With ASA
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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kjær, J.
Right arrow Articles by Grant, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kjær, J.
Right arrow Articles by Grant, R.
Agricola
Right arrow Articles by Kjær, J.
Right arrow Articles by Grant, R.
Related Collections
Right arrow Water Quality
Right arrow Agricultural Pesticides
Right arrow Field-Scale Studies
Right arrow Pesticides
Right arrow Vadose Zone Processes and Chemical Transport
Published in J. Environ. Qual. 34:608-620 (2005).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Vadose Zone Processes and Chemical Transport

Leaching of Glyphosate and Amino-Methylphosphonic Acid from Danish Agricultural Field Sites

Jeanne Kjæra,*, Preben Olsenb, Marlene Ulluma and Ruth Grantc

a Geological Survey of Denmark and Greenland, Øster Vold 10, DK-1350 Copenhagen, Denmark
b Danish Institute of Agricultural Sciences, Research Centre Foulum, DK-8830 Tjele, Denmark
c National Environmental Research Institute, Vejlsøvej 25, DK-8600 Silkeborg, Denmark

* Corresponding author (jkj{at}geus.dk)

Received for publication November 12, 2003.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Pesticide leaching is an important process with respect to contamination risk to the aquatic environment. The risk of leaching was thus evaluated for glyphosate (N-phosphonomethyl-glycine) and its degradation product AMPA (amino-methylphosphonic acid) under field conditions at one sandy and two loamy sites. Over a 2-yr period, tile-drainage water, ground water, and soil water were sampled and analyzed for pesticides. At a sandy site, the strong soil sorption capacity and lack of macropores seemed to prevent leaching of both glyphosate and AMPA. At one loamy site, which received low precipitation with little intensity, the residence time within the root zone seemed sufficient to prevent leaching of glyphosate, probably due to degradation and sorption. Minor leaching of AMPA was observed at this site, although the concentration was generally low, being on the order of 0.05 µg L–1 or less. At another loamy site, however, glyphosate and AMPA leached from the root zone into the tile drains (1 m below ground surface [BGS]) in average concentrations exceeding 0.1 µg L–1, which is the EU threshold value for drinking water. The leaching of glyphosate was mainly governed by pronounced macropore flow occurring within the first months after application. AMPA was frequently detected more than 1.5 yr after application, thus indicating a minor release and limited degradation capacity within the soil. Leaching has so far been confined to the depth of the tile drains, and the pesticides have rarely been detected in monitoring screens located at lower depths. This study suggests that as both glyphosate and AMPA can leach through structured soils, they thereby pose a potential risk to the aquatic environment.

Abbreviations: AMPA, amino-methylphosphonic acid • BGS, below ground surface


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
GLYPHOSATE is one of the most used herbicides worldwide. In 1997, sales of the active ingredient glyphosate accounted for 29.4% of all herbicides sold in Denmark. By 2003, this had increased to 42.7% (Danish Environmental Protection Agency, 2004). The leaching risk of glyphosate is generally regarded as being very low. Several studies have shown that glyphosate is degradable and strongly adsorbed to topsoils (e.g., Rueppel et al., 1977; de Jonge and de Jonge, 1999; Aamand and Jacobsen, 2001). The USEPA (1993) has reported Kd values of glyphosate in the range between 62 and 175 L kg–1 whereas Aamand and Jacobsen (2001) reported 65 to 147 L kg–1 for Danish conditions. In a review by Geisy et al. (2000), DT50 values (i.e., dissipation time for 50% of added pesticide) of glyphosate have been summarized for a variety of soil types. They indicate DT50 values as determined in laboratories to be less than 60 d, while those generated by field dissipation studies range between 1 and 197 d, with arithmetic and geometric mean values of 32 and 17 d, respectively. Moreover, laboratory studies of glyphosate transport in either repacked soil columns or by thin-layer chromatography all suggest that the leaching risk in soils is low (Sprankle et al., 1975; Roy et al., 1989; Cheah et al., 1997). The supposition that the leaching risk in Danish soils is low is supported by a lysimeter study on a sandy loam soil (Fomsgaard et al., 2003) and a study of glyphosate transport in 20-cm undisturbed sandy topsoil columns (de Jonge et al., 2000). In the latter study, however, glyphosate was found to leach from the top of a loamy soil under conditions of intense, simulated precipitation occurring shortly after application.

Most reported data on glyphosate transport in soil derive from laboratory and lysimeter studies and provide little, if any, information on the inherent variability of the soil parameters affecting leaching at the larger scale of fields. This is of particular importance for structured soils, where preferential flow can have a major impact on leaching. In fact, various field studies suggest that considerable preferential transport of several pesticides occurs to a depth of 1 m (e.g., Brown et al., 1995; Flury, 1996; von Kördel et al., 1997; Zehe and Flühler, 2001; Kladivko et al., 2001; Petersen et al., 2002). Controlled experiments analyzing glyphosate leaching under field conditions are limited. The present study thus examined glyphosate transport to determine the risk that glyphosate might leach under Danish field conditions. Among the various routes of dissipation, this study focuses on the leaching processes, relating its dynamic to soil and hydrological properties.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The field study was conducted within the framework of the Danish Pesticide Leaching Assessment Programme (PLAP). The monitoring program focuses on pesticides used in crop production, and intensively monitors leaching from six different agricultural test fields representative of Danish conditions (Lindhardt et al., 2001; Kjær et al., 2002, 2003). The PLAP is intended to serve as an early warning system, providing decision makers with advance warning if approved pesticides, when used in accordance with the regulations, leach in unacceptable concentrations.

Site Description
The sites on which glyphosate was applied comprised a coarse sandy soil at Jyndevad and two loamy soils at Estrup and Faardrup (Fig. 1) . Jyndevad and Estrup are practically flat with slopes less than 2%, whereas the terrain at Faardrup slopes gently by 2 to 5%. The loamy sites have a tile drain system at an average depth of about 1 m. All three sites are characterized by a relatively shallow water table located 1 to 3 m BGS. At each site, three soil profiles were described to a depth of 1.5 to 1.9 m, always including the C horizon. Important parameters for a selection of profiles are given in Table 1. Moreover, a three-dimensional geological model, based on geological information from various boreholes profiles, was established for each of the three test sites (Fig. 2) .



View larger version (76K):
[in this window]
[in a new window]
  		Fig. 1. Location of the three test sites Jyndevad, Estrup, and Faardrup.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Physical and chemical properties{dagger} of selected soil profiles{ddagger}.

 


View larger version (50K):
[in this window]
[in a new window]
  		Fig. 2. Geological models for the three test sites Jyndevad, Estrup, and Faardrup. White arrows indicate the north direction, and black dots and vertical black lines the positions and depth of those boreholes providing the geological information by Lindhardt et al. (2001).

 
The aquifer at Jyndevad consists of meltwater sand with local occurrences of thin clay and silt beds. Out of three soil profiles described at Jyndevad one was classified according to the USDA Soil Taxonomy as Arenic Eutrudept and two as Humic Psammentic Dystrudepts. One of the latter is included in Table 1. Except for the Bhs horizon, having a weak, coarse subangular structure due to deposition of humus, aluminium, and iron, the soil is almost structureless owing to the coarse grain size distribution.

The Faardrup aquifer consists of clayey till with local occurrences of small channels and basins containing meltwater clay and sand. The three soil profiles at Faardrup were classified according to USDA Soil Taxonomy as Haplic Vermudoll, Oxyaquic Hapludoll, and Oxyaquic Argiudoll, of which only the latter is included in Table 1. Signs of clay eluviation could be seen all over the profile as thick clay coatings on aggregate surfaces and wormholes. There were clear pseudogley stripes, indicative of temporary stagnant water in moist periods. Free chalk was present in the profile.

Estrup is located in central Jutland, west of the Main Stationary Line on a glacial moraine of Saalian Age. Estrup has thus been exposed to weathering, erosion, leaching, and other geomorphologic processes for a much longer period (about 90000 yr) than that of the other two sites (Houmark-Nielsen, 1987) dating back to the Weichselian glaciation some 10000 yr ago. Compared with Faardrup and Jyndevad, Estrup is highly heterogeneous with considerable variation in both topsoil and aquifer characteristics (Table 1 and Fig. 2). Such heterogeneity is quite common for this geological formation, however. The geological structure is complex comprising a clay till core with deposits of different age and composition. The three profiles were classified as Aquic Argiudoll, Abruptic Argiudoll, and Fragiaquic Glossudalf, the latter two of which are included in Table 1. Numerous macropores and sand-filled fissures that range in size from a few millimeters to tens of centimeters were observed at the site along with heavy calcareous clay near the surface. The latter is unusual for soils of the Saale, usually being deeply leached.

Agricultural Management
Cultivation of the three sites is in line with conventional agricultural practices applied in the regions, whereas pesticides are applied in the maximal permitted dose in accordance with the regulations. Glyphosate was applied once at Jyndevad (22 Sept. 1999) and Estrup (13 Oct. 2000) and twice at Faardrup (11 Aug. 1999 and 14 Oct. 2000). At Estrup, the product used was Roundup Bio (Monsanto, St. Louis, MO) at a rate of 4 L ha–1, equivalent to 1.44 kg glyphosate ha–1. At Faardrup and Jyndevad, 2.0 L ha–1 of Roundup 2000 was applied, equivalent to 0.8 kg glyphosate ha–1 (Table 2). To describe water transport and especially to assure that the water being sampled had infiltrated on the test field, a bromide tracer (30 kg KBr ha–1) was also applied to each field.


View this table:
[in this window]
[in a new window]
  		Table 2. Management practice at the three test sites.

 
Monitoring
To avoid unintended leaching of pesticide due to the installation and presence in the ground of sampling equipment, all installations and soil sampling deeper than 20 cm were restricted to a buffer zone surrounding the treated area (Fig. 3) . Suction cups and monitoring wells were established in June 1999 at Faardrup and Jyndevad and in November 1999 at Estrup. The drainage system of Faardrup was established in 1944 and that of Estrup before 1965. At the loamy sites, the drainage system was modified before the monitoring such that tile-drainage water was only collected from the treated area and directed into a newly established well fitted with a Thomson weir (30° V-notch). The water head behind the weir was measured automatically using a pressure transducer (PDCR1830; Druck, Leicester, UK) coupled to a CR10X data logger (Campbell Scientific, Logan, UT). A tipping bucket rain gauge system was used for measurements of precipitation at each of the three sites. Tile-drainage water, ground water, and soil water sampled in the unsaturated zone were analyzed for bromide and pesticides.



View larger version (43K):
[in this window]
[in a new window]
  		Fig. 3. Overview of the Estrup test site, illustrating the typical layout of monitoring devices at a tile-drained Pesticide Leaching Assessment Programme (PLAP) site. The innermost area indicates the cultivated area, while the gray area indicates the surrounding buffer zone. The positions of the various installations are indicated, as is the direction of the ground water flow (by an arrow).

 
Soil water samples were collected monthly using 16 Teflon suction cups (Prenart, Frederiksberg, Denmark), each connected via a single length of PTFE tubing to a sampling bottle placed in a refrigerator in the instrument shed. The soil water was extracted by applying a continuous vacuum (of about 80 KPa) to each of the suction cups one week before sampling. The 16 suction cups were clustered in four groups. Two groups were installed respectively 1 and 2 m BGS at location S1, and the remaining two groups were installed respectively 1 and 2 m BGS at location S2 (Fig. 3). Each group of suction cups thus consisted of four individual cups covering a horizontal distance of 2 m. Chemical analysis was performed on a single, pooled water sample for each of the four groups.

Ground water samples were collected monthly from several vertical and horizontal monitoring wells (Fig. 3). Each vertical monitoring well, installed in the surrounding buffer zone, consists of four 1-m screens covering the upper approximately 4 m of the saturated zone. In addition, horizontal monitoring wells were installed 3.5 m beneath the two loamy test sites; two at Faardrup and one at Estrup. Each horizontal monitoring well consists of 18-m screens providing integrated water samples characterizing ground water quality just beneath the test site (Fig. 3). The horizontal screens are installed by drilling from the buffer zone on the one side of the field to the buffer zone on the opposite side, without causing any disturbance to the topsoil within the cultivated area. Sampling of horizontal and vertical monitoring wells was performed using a whale pump (permanently installed in each screen) and a peristaltic pump, respectively. At the sandy site (Jyndevad), each well was purged by removing a volume of water equivalent to three times the volume of the saturated part of the well before water sampling. At the loamy sites Estrup and Faardrup, the wells were purged by emptying them the day before sampling.

Drainage water samples were collected by means of time-proportional and flow-proportional sampling at the loamy sites as described by Plauborg et al. (2003). Time-proportional sampling refers to sampling at regular intervals throughout the whole drainage season. During the period of continuous drainage runoff, a 70-mL subsample was collected every hour independently of the flow rate. Twenty-four samples were collected per bottle giving 1680 mL d–1. Chemical analysis was then performed on a weekly basis on a pooled sample, derived from the seven bottles. Flow-proportional sampling refers to sampling drainage runoff induced by sudden precipitation events. The flow-proportional sampler was only activated during storm events, with sampling being performed for 1 to 2 d depending on the intensity of the event. Hence, each flow event was activated by a predefined rise in water level and runoff within the preceding 12-h period. Sampling was controlled by the flow rate. The collection of each sample began when the accumulated flow rate exceeded a predefined level, depending on the month of the year. Levels of predefined rise and accumulated flow rate were set or adjusted individually for each site by experience (Plauborg et al., 2003). Each subsample was 200 mL, yielding nine subsamples per bottle and a maximum of 72 subsamples per flow event. To obtain a weighted average concentration for each storm event, the chemical analysis was then performed on each pooled water sample.

The weighted average concentration of pesticides in the tile-drainage water was subsequently calculated according to the equation:


where n is number of weeks within the period of continuous drainage runoff; Vi is weekly accumulated tile-drainage water (mm wk–1); Vfi is tile-drainage water accumulated during a "flow event" (mm per storm event); Cfi is pesticide concentration in the "event samples" collected by means of the flow-proportional sampler (µg L–1); and Cti is pesticide concentration in the weekly samples collected by means of the time-proportional sampler (µg L–1).

Methods of Analysis
The pesticide analyses were all performed in commercial laboratories on decanted water samples. The water samples were preserved in the laboratory by adjusting to pH 2. A sample of 500 mL was concentrated on a column of Chelex 100. After washing with 0.1 M HCl the analytes were eluted with 6 M HCl. The eluate was further cleaned on a column of AG 1-X8. The eluate was then treated with trifluoroacetic anhydride and 2,2,3,3,4,4,4-heptafluoro-1-butanol to derivatize the analytes. The derivates of glyphosate and AMPA were measured by gas chromatography–mass spectrometry (GC–MS) using a GC column of 5% phenyl methylsiloxane and the MS in electron impact (EI) mode. The mass to charge ratios (m/z) 612, 611, and 584 were used for identification of the glyphosate derivative and the m/z 446, 372, and 502 were used for the AMPA derivative. Calibration was performed on standards taken through the whole procedure, using internal standard calculations. As part of the internal quality control, two replicates of control samples were included in every series of samples. The control samples had a concentration level of 0.05 µg L–1. At this level the method recovery was 106 to 109% for glyphosate and 104 to 107% for AMPA, whereas coefficient of variation was 12% for both glyphosate and AMPA. The limit of detection was determined to be below 0.01 µg L–1 for both compounds. Blank samples consisting of high performance liquid chromatography (HPLC) water were sent to the laboratory once a month together with the ordinary ground water samples. Blank samples were labeled with coded reference numbers so that the laboratories were unaware of which samples were controls and blanks. No pesticides were detected in blank samples, thus indicating that no contamination of the samples occurred in the laboratory.

Water Balances
The monitoring data were supported by hydrological modeling (MACRO Version 4.3; Jarvis, 2001) providing an overall water balance and a description of the soil water dynamics in the unsaturated zone. The MACRO model was applied to each site, covering the soil profile to a depth of 5 m BGS and always including the water table. The model was parameterized using measured data or literature or default values. Model performance was evaluated by comparing simulated and measured data. The latter comprised measurements of tile-drainage water, the location of the water table determined using piezometers located in the buffer zone, and the soil water content determined using time-domain reflectometry probes placed at three depths (25, 60, and 110 cm BGS) within the two profiles S1 and S2 (see Fig. 3). The resulting water balances for the three sites are summarized in Table 3. For a detailed description of data acquisition, model setup, and model performance, see Kjær et al. (2002)(2003).


View this table:
[in this window]
[in a new window]
  		Table 3. Annual water balances for Jyndevad, Estrup, and Faardrup.{dagger}

 

   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Estrup
Glyphosate and its degradation product AMPA leached from the root zone, entering the tile drains (1 m BGS) in average concentrations exceeding 0.1 µg L–1, especially in the case of glyphosate. Thus the average concentration of glyphosate in the drainage water during the 2000–2001 leaching period (31 Oct. 2000 to 8 May 2001) was 0.54 µg L–1, while that of AMPA was 0.17 µg L–1 (Table 4, Fig. 5B and 5C). Annual average concentration during the first year after application (26 Oct. 2000–25 Oct. 2001) was 0.41 µg L–1 for glyphosate and 0.14 µg L–1 for AMPA. It should be noted that analytical data below the detection limit did not affect the estimated average concentration in Table 4. Applying a concentration equal to either zero or the detection limit in the case of samples for which the concentration was below the limit of detection, was thus found to have no impact on the estimated concentration.


View this table:
[in this window]
[in a new window]
  		Table 4. Glyphosate and AMPA (amino-methylphosphonic acid) in tile-drainage water at Estrup during the two monitoring years.{dagger}

 


View larger version (38K):
[in this window]
[in a new window]
  		Fig. 5. Precipitation (A) together with concentration of glyphosate (B), AMPA (amino-methylphosphonic acid) (C), and bromide (D) in the tile-drainage water in 2000–2001 at Estrup. The gray vertical lines indicate the date of application.

 
Leaching of glyphosate appeared to be governed by the pronounced macropore flow characterizing the Estrup site. Evidence of macropore flow at Estrup is provided by soil hydraulic properties and drainage water dynamics. Saturated hydraulic conductivity has been measured in the laboratory using both small (6.1-cm diameter, 100 cm3) and large (20-cm diameter, 6280 cm3) soil samples (Table 1). Conductivity was found to increase considerably with increasing sample size, indicating the presence of macropore flow in the more representative large soil samples as also suggested by Iversen et al. (2001). Similarly, measured hydraulic characteristics in A and B horizons showed a marked increase of the conductivity when approaching full saturation, also indicating a high degree of preferential flow through macropores when the soil is fully saturated (data not shown, see Lindhardt et al., 2001). Moreover, the drainage runoff, being very sensitive, responding within a few hours even to small rain events (Fig. 4) , also indicates a flow regime dominated by macropore flow bypassing the bulk of the soil. The supposition of macropore flow was also supported by field observations revealing numerous macropores and sand-filled fissures, all of which could facilitate macropore transport.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 4. Hourly precipitation and tile-drainage water at two storm events occurring at Estrup shortly after application of glyphosate in October 2000.

 
Glyphosate (1.44 kg ha–1) was applied on stubble on 13 Oct. 2000. No precipitation fell in the subsequent 10-d period before plowing on 23 Oct. 2000. Thereafter, drainage runoff responded rapidly to the first storm events. The heavy storm events in October and November 2000 induced marked, rapid leaching of both glyphosate and AMPA in concentrations reaching 2.1 and 0.73 µg L–1, respectively (Fig. 5B and Fig. 5C). When comparing the mass leached during the storm event with the total mass being leached during the entire drainage period it could be seen that the transport of the pesticides was event-driven. The 11 storm events that activated the flow-proportional sampler accounted for 94% of the glyphosate and 93% of the AMPA leached during the drainage period 2000–2001. This indicates fast transport of water and pesticide, presumably through macropores, a finding consistent with several other field studies indicating that macropore transport can induce strongly sorbed chemicals to move rapidly through soil [for reviews see Flury (1996) and Kladivko et al. (2001)]. Rapid macropore transport of pendimethalin (a strongly sorbed compound) was observed by Petersen et al. (2002) and Traub-Eberhard et al. (1995) during leaching to tile drains in sandy loam soils of Denmark and Germany. Further, Nilsson et al. (2000) observed preferential transport of both AMPA and glyphosate in a Danish, glacial till. Glyphosate and AMPA were detected in the drainage water 19 d after application in concentrations reaching 11.0 and 0.6 µg L–1, respectively. Evidence of macropore transport causing relatively high glyphosate leaching rates in topsoils was also provided by de Jonge et al. (2000) in a laboratory study of glyphosate transport using 20-cm undisturbed sandy loam topsoil columns. About 11% of the applied dosage leached following 20 mm of precipitation applied immediately after pesticide application. A 4-d time lag between application and the 20-mm rain event reduced total leaching to only 0.6 to 2% of the applied dosage.

The bromide leaching pattern was somewhat different than that of glyphosate. During the first drainage period (2000–2001), the concentrations of glyphosate attained during storm events (activating the flow-proportional sampler) were considerably higher than those attained during continuous drainage runoff (Fig. 5B). With bromide, however, the concentrations reached during storm events were similar to those reached during continuous drainage runoff (Fig. 5D). The 11 storm events accounted for only 54% of the bromide leached during the drainage period of 2000–2001. Compared with glyphosate, macropore transport thus had a minor impact on bromide leaching, much more of which took place through the soil matrix. A likely explanation for this could be differences in sorption and diffusion characteristics. Compared with glyphosate, the residence time of bromide in the root zone was much longer before the autumn storm events, due to the earlier application of the bromide (see Table 2). The longer residence time and the higher diffusion coefficient of bromide (Mortensen et al., 2004) thus allowed a greater proportion of the bromide to enter the soil matrix, where it was unaffected by the bypass flow in the macropores. If a significant proportion of the precipitation flows into the macropores at the soil surface, it would have little contact with the soil matrix. The majority of the bromide present in the soil matrix would thus be "protected" from bypass flow, as also suggested by Larsson and Jarvis (1999). Unlike glyphosate and AMPA, bromide will not be retained by sorption. Although "protected" from bypass flow, bromide located in the soil matrix is still prone to leaching by water infiltrating through the soil matrix.

Both AMPA and glyphosate leached continuously throughout the whole 6-mo drainage runoff period of 2000–2001. Although the level of concentration was much lower, leaching of AMPA in particular continued during the subsequent drainage runoff period of 2001–2002 (Table 4 and Fig. 6) . The heavy storm events occurring in February and March 2002, in particular (i.e., 1.5 yr after application), induced leaching of AMPA in concentrations reaching 0.15 µg L–1 (Fig. 6C).



View larger version (40K):
[in this window]
[in a new window]
  		Fig. 6. Precipitation (A) together with concentration of glyphosate (B), AMPA (amino-methylphosphonic acid) (C), and bromide (D) in the tile-drainage water in 2001–2002 at Estrup. Bromide and glyphosate were applied in April 2000 and October 2000, respectively. Open diamonds and triangles in indicate that the concentrations were below the detection limit of either 0.01 µg L–1 for AMPA and glyphosate or 0.1 mg L–1 for bromide.

 
The difference in leaching pattern of glyphosate and AMPA is further illustrated in Fig. 7 . With glyphosate, leaching was mainly confined to the first four months after application and was ascribed to single rain events. Leaching decreased with time, and the amount leached during 2001–2002 was negligible. The leaching pattern of AMPA differs from that of glyphosate in that minor leaching occurred over a much longer time period. The amount of AMPA leached was about 60% lower than that of glyphosate. Leached mass also decreased with time, but unlike with glyphosate, leaching continued throughout the second leaching period. The decreased leaching of glyphosate is likely to be attributable to the glyphosate being either sorbed or degraded into AMPA. The latter is illustrated by the fact that AMPA was detected in both the very first water samples and in an increasing fraction with time (Table 4).



View larger version (18K):
[in this window]
[in a new window]
  		Fig. 7. Accumulated leached mass (% of applied glyphosate) of glyphosate and AMPA (amino-methylphosphonic acid) together with accumulated tile-drainage water (secondary axis) during the two leaching periods at Estrup. Leached mass of AMPA is expressed as glyphosate equivalent and the black vertical line indicates the date of application.

 
That leaching of AMPA occurs a relatively long time after application is indicative of (i) AMPA being retained in the soil for a rather long time before further degradation, and (ii) a minor release occurring from the uppermost meter of the soil. Experimental data supporting whether this minor release was due to either a slow degradation of soil-adsorbed glyphosate or a gradual desorption of AMPA from soil are not available. However, the results of Jacobsen (2003) suggest the latter to be a likely explanation. Jacobsen (2003) thus observed a similar tendency when evaluating the degradation, sorption, and persistence of glyphosate and AMPA in a fractured clay soil profile in Denmark. Here glyphosate degraded relatively rapid into AMPA. This was, however, retained in the soil for a rather long period of time. Thus, more than two years after the application, up to 26% of the applied glyphosate was still retained in the soil, predominantly in the form of AMPA (>95%), mainly in the plow layer. Jacobsen (2003) found that AMPA had the highest sorption potential, and suggested that the slow desorption of AMPA was the limiting factor for the total mineralization of glyphosate.

During the two monitored years, 0.12 and 0.09% of the applied dosage leached as glyphosate and AMPA, respectively (Fig. 7). The mass leached is obviously highly site-specific, depending on the soil properties and hydrological conditions following the pesticide application. As far as we are aware, quantitative field data on glyphosate leaching have not been reported. Nevertheless, this rather small mass loss was within the range of several other field pesticide leaching studies. Thus, in a recent review, Kladivko et al. (2001) suggested the mass of a large range of herbicides lost to subsurface drains to be only 0.5% of the amount applied, and often less than 0.1%.

Evidence of glyphosate and AMPA leaching has hitherto only been seen in the tile-drainage water. Apart from three samples containing 0.01 to 0.04 µg L–1 glyphosate, neither AMPA nor glyphosate have yet been detected in samples obtained from suction cups or ground water monitoring screens located beneath the drainage system. The bulk of the leached glyphosate and AMPA probably left the system with tile-drainage water, since the water balance suggests that the predominant part (64–71%) of the percolation ran off through the drainage system (Table 3). During the two monitoring years (1 July 2000–30 June 2002), ground water mean recharge was 213 mm yr–1, corresponding to 33% of the percolation. The elevated bromide concentration in monitoring screens located beneath the drainage system also indicates the occurrence of ground water recharge at the Estrup site (data not shown; see Kjær et al., 2003). Due to decreased hydraulic conductivity, water and solute transport at Estrup were much slower beneath the drainage system than above it, however (Lindhardt et al., 2001). The longer transit time allows for dispersion, dilution, sorption, and degradation (Haria et al., 2003). The rare detection in the monitoring screens located beneath the drainage system could thus be due to these processes reducing further transport.

The reason why neither AMPA nor glyphosate have yet been detected in the suction cups might be due to the differences in the sampling method. The suction cups primarily extract water from the soil matrix (Schoen et al., 1999; Magid et al., 1992). Moreover, the water sampled this way is only representative of a relatively small part of the test site, whereas the drainage system provides integrated samples, capturing water infiltrating through both the soil matrix and the macropores. The high proportion of leaching mediated by macropore flow is therefore more likely to be detected in water samples from the drainage system. Finally, monthly sampling on this structured soil may be too infrequent for capturing the rapid water transport following a major rainstorm.

When evaluating leaching at Estrup, it should be noted that the climatic conditions during the monitoring period were not exceptional. October and November 2000 were characterized by high precipitation input, exceeding the monthly normal by 20%, and heavy storm events reaching a maximum of 30 mm d–1. This precipitation pattern, in terms of daily and monthly precipitation, is not unusual for the Estrup region, however. Similar patterns have occurred in the preceding 10 yr (Kjær et al., 2002).

Jyndevad
Leaching of glyphosate did not occur at Jyndevad. At Jyndevad, 0.8 kg ha–1 of glyphosate was applied to stubble on 22 Sept. 1999. Eight days later a heavy storm event (43 mm d–1) initiated percolation, as estimated by the MACRO model (Fig. 8D) . During the following October, precipitation amounted to 115 mm. The precipitation pattern that followed the glyphosate application at Jyndevad was similar to that of Estrup (Fig. 8A and 8D), whereas the time lag between pesticide application and the onset of percolation was even shorter at Jyndevad. Nevertheless the short time lag at Jyndevad appeared to be sufficient to prevent leaching. Apart from three ground water samples containing 0.01 to 0.02 µg L–1 of AMPA, glyphosate and AMPA have not yet been detected in any of the water samples analyzed. A plausible explanation could be that there are no macropores in this unstructured, coarse, sandy soil. No sign of macropore flow was found in measured hydraulic conductivities, which were independent of sample size (Table 1). Similarly, measured retention characteristics showed no marked increase of the conductivity when approaching full saturation (data not shown, see Lindhardt et al., 2001). The lack of macropores provides much better conditions for sorption and/or degradation processes due to a longer residence time in the root zone, as well as better contact between the infiltrating water and the surrounding soil matrix.



View larger version (57K):
[in this window]
[in a new window]
  		Fig. 8. Precipitation (hanging bars on primary axis), tile-drainage water, and percolation (secondary axis) following the application of glyphosate at the test sites. Percolation was estimated by means of Macro 4.2 (Kjær et al. 2003). The light gray vertical lines indicate the dates of application.

 
Moreover, the elevated Fe and Al content in the upper soil layer (Table 1) might provide sufficient sorption capacity to prevent leaching of glyphosate. Aluminum and Fe are suggested to be important sorbents for glyphosate in soils (Piccolo et al., 1994; Gerritse et al., 1996; de Jonge et al., 2001; Jacobsen, 2003). De Jonge et al. (2001) analyzed sorption isotherms of soil samples from Jyndevad using the modified Freundlich model (Kmf) of Sibbesen (1981). The Kmf values ranged from 119.8 to 136.6 for soil samples with soil properties very similar to the one reported in this study. According to Table 1 measured amounts of Fe and Al in this soil type did not deviate much from that of Estrup and Faardrup, thus one could then expect similar high sorption capacity in these loamy soils. Measurements, however, were in all cases made on bulk samples. Rasmussen et al. (2001), on a structured soil (Typic Agriudalf), found that coatings in fractures and macropores contained less Fe and Al oxides than the soil horizons as such. On this structured soil, they demonstrated that when all flow went through macropores the soil would have five times less sorption capacity for P, competing for the same sorption sites as glyphosate (Sprankle et al., 1975), than when all flow passed through the matrix. Aluminum and Fe as determined from bulk samples may give a fair idea of sorption capacity on structureless soil, but may be misleading on a structured soil.

Field data on the leaching of glyphosate in sandy soils seem not to have been reported in the literature. The supposition that the risk of glyphosate leaching in sandy soils is low is consistent with the laboratory studies of de Jonge et al. (2000); despite extreme application of water and a short time lag between application of glyphosate and the subsequent precipitation events, the active pore volume of the coarse sandy soil was large enough to ensure effective sorption of glyphosate, resulting in very low glyphosate concentrations in the effluent water.

Faardrup
Glyphosate leaching was minor at Faardrup as it was only found on two occasions in four water samples from the drainage system (time-proportional and flow-proportional) and in three samples from the ground water monitoring wells. The concentration range was 0.01 to 0.093 µg L–1.

The measured soil properties showed that Faardrup had fissures and macropores similar to that of Estrup. Similar to Estrup, hydraulic conductivity increased with increasing sample size (Table 1) and measured conductivity increased markedly when approaching full saturation (data not shown; see Lindhardt et al., 2001). Although the Faardrup soils had the necessary fissures and macropores, precipitation input as well as intensity was low, allowing the substances to reside in the root zone long enough for processes of sorption and degradation and thereby the prevention of glyphosate leaching. At Faardrup, both applications, occurring August 1999 and October 2000, were thus followed by moderate precipitation input, and percolation commenced more than 1.5 mo after glyphosate application (Fig. 8B and 8C). According to a review of Geisy et al. (2000), DT50 values (i.e., dissipation time for 50% of added pesticide) generated by field dissipation studies ranged between 1 and 197 d with arithmetic and geometric mean values of 32 and 17 d, respectively.

AMPA was detected more frequently, however, being found in 10 samples from tiles, six from suction cups, and two from ground water screens. AMPA was first detected in suction cups in April 2001, 5 mo after the last application. From May 2001 to January 2002, AMPA was detected at low concentrations (0.02–0.11 µg L–1) in a few drainage water samples (Fig. 9) . It was last detected in February 2002 in a sample from the vertical monitoring well. Since glyphosate was applied twice, it is not possible to relate the findings to one specific application. The more frequent detection of AMPA, a relatively long time after glyphosate application, is in line with the findings at Estrup, indicating minor release occurring in the uppermost part of the soil system.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 9. Precipitation (A) together with concentration of AMPA (amino-methylphosphonic acid) (B) in the tile-drainage water in 2001–2002 at Faardrup. Glyphosate was applied in May 1999 and October 2000. Open diamonds and triangles indicate that the concentrations were below the detection limit of 0.01 µg L–1.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
This study suggests that as both glyphosate and AMPA can leach through structured soils, they thereby pose a potential risk to the aquatic environment. In the loamy soil at Estrup, glyphosate and AMPA thus leached from the root zone into the tile drains (1 m BGS) in average concentrations exceeding 0.1 µg L–1, which is the EU threshold value for drinking water. Leaching of glyphosate and AMPA was confined to the depth of the tile drains, and both compounds were rarely detected in the monitoring screens situated beneath the tile drains.

On the other loamy soil, Faardrup, where precipitation was less and of lesser intensity, the residence time in the root zone seems to have been sufficient to prevent leaching of glyphosate. Although leaching of AMPA was observed at this site, the concentrations were generally low, being on the order of 0.05 µg L–1 or less. On the coarse sandy soil, however, the risk of glyphosate leaching was found to be negligible. Infiltrating water passed through a matrix rich in Al and Fe, thus providing good conditions for both sorption and degradation.


   ACKNOWLEDGMENTS
 
The authors thank the many people who have contributed to this work over the years including establishment of field sites and monitoring design by Bo Lindhardt, Christian Abildtrup, Henrik Vosgerau, Bo Vangsø Iversen, Henning Hougaard, Søren Torp, and Finn Plauborg, and ongoing field monitoring and data preparation by Peter Carlsen, Jens Barsballe, Verner Hansen, Birgit Sørensen, Pia G. Jensen, Carsten B. Nielsen, Carl H. Hansen, Søren Nielsen, and Lasse Gudmundsson. The Danish Pesticide Leaching Assessment Programme funded the work.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2005 34: 403-407. [Full Text]  



This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
A. L. Cerdeira and S. O. Duke
The Current Status and Environmental Impacts of Glyphosate-Resistant Crops: A Review
J. Environ. Qual., August 9, 2006; 35(5): 1633 - 1658.
[Abstract] [Full Text] [PDF]

Home page
 J. Environ. Qual.Home page
W. W. Stone and J. T. Wilson
Preferential Flow Estimates to an Agricultural Tile Drain with Implications for Glyphosate Transport
J. Environ. Qual., August 9, 2006; 35(5): 1825 - 1835.
[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 Related articles in JEQ
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kjær, J.
Right arrow Articles by Grant, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kjær, J.
Right arrow Articles by Grant, R.
Agricola
Right arrow Articles by Kjær, J.
Right arrow Articles by Grant, R.
Related Collections
Right arrow Water Quality
Right arrow Agricultural Pesticides
Right arrow Field-Scale Studies
Right arrow Pesticides
Right arrow Vadose Zone Processes and Chemical Transport

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