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


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow 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 ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by van Alphen, B. J.
Right arrow Articles by Stoorvogel, J. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by van Alphen, B. J.
Right arrow Articles by Stoorvogel, J. J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by van Alphen, B. J.
Right arrow Articles by Stoorvogel, J. J.
Related Collections
Right arrow Spatial Distribution
Right arrow Ground Water Quality
Right arrow Best Management Practices
Right arrow Agricultural Pesticides
Right arrow Water Pollution
Journal of Environmental Quality 31:797-805 (2002)
© 2002 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORTS
Ground Water Quality

Effects of Soil Variability and Weather Conditions on Pesticide Leaching— A Farm-Level Evaluation

B. J. van Alphen and J. J. Stoorvogel*

Laboratory of Soil Science and Geology, Wageningen University, P.O. Box 37, 6700 AA Wageningen, the Netherlands

* Corresponding author (jetse.stoorvogel{at}bodlan.beng.wag-ur.nl)

Received for publication April 9, 2001.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
In line with European regulations, Dutch law imposes an environmental threshold of 0.1 µg L-1 on pesticide concentrations in ground water. During registration, the risk of exceeding this threshold is assessed through simulations for one or a few standard scenarios that do not reflect spatial variability under field conditions. The introduction of precision agriculture, where soil variability is actively managed, can increase control over pesticide leaching. This study presents a step-wise evaluation of the effects of soil variability and weather conditions on pesticide leaching. The evaluation was conducted on a 100-ha arable farm and aimed at identifying opportunities for precision management. As a first step, a relative risk assessment identified pesticides presenting a relatively high risk to the environment. Second, the effect of weather conditions was analyzed through 20 years of simulations for three distinct soil profiles. Results were summarized in cumulative probability plots to provide a probabilistic characterization of historical weather data. The year matching 90% probability (1981) served as a reference to simulate pesticide leaching from 612 soil profiles. After interpolation, areas where concentrations exceeded the environmental threshold were identified. Out of a total of 19 pesticides, isoproturon [N-dimethyl-N'-(4-(1-methylethyl)phenyl)urea], metribuzin [4-amino-6-tert-butyl-3-(methylthio)-as-triazin-5(4H)-one], and bentazon [2,1,3-benzothiadiazin-4(3H)-one, 3-isopropyl-, 2,2-dioxide] showed the highest risk for leaching. Leaching was strongly affected by soil variability at the within-field, field, and farm levels. Opportunities for precision management were apparent, but depended on the scale level at which environmental thresholds were implemented. When legislation is formulated in this issue, the presented step-wise evaluation can serve as a basis for identification and precision management of high-risk pesticides.

Abbreviations: DT50, degradation time • EU, European Union • Kom, coefficient of sorption to organic matter • SOM, soil organic matter


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
IN THE LATE 1980s, Leistra and Boesten (1989) found pesticide residues in ground water bodies across Europe. An extensive review by Ritter (1990) reports similar results for the Unites States. In the Netherlands, ground water provides an important source of drinking water and its protection is assigned high priority. To improve ground water quality as well as the quality of natural resources in general, the Dutch government introduced three environmental criteria against which plant protection products must be tested for registration (Health Council of the Netherlands, 2000). These criteria impose limits on persistence in soil (degradation time [DT50] <90 d), leaching into the ground water (concentration in ground water <0.1 µg L-1), and the risk to aquatic organisms (limits on peak concentrations in surface waters based on toxicity levels for fish, Daphnia, and algae). These criteria are in line with the registration directive (91/414/EEC) of the European Union (EU).

An efficient means of assessing pesticide leaching into ground water is provided by simulation models (Vanclooster et al., 2000). As a result, environmental fate modeling now plays an important role in the EU registration process (FOCUS Groundwater Scenarios Workgroup, 2000). However, problems are encountered when modeling results, which are almost exclusively attained at the point level, are extrapolated to areas. Extrapolation is particularly difficult since pesticide persistence and adsorption have been shown to present great spatial and temporal variation (e.g., Walker et al., 2000). Stoorvogel et al. (1999) simulated nematicide leaching from more than 400 soil profiles located on a banana (Musa AAA) plantation in Costa Rica. Leaching proved to be confined to small areas (so-called hot spots) and to particular periods of the year. Spatial variability was large and could not be described adequately using representative soil profiles for each soil type.

The above illustrates that simulations for one or a few standard scenarios can result in large overestimation of risks associated with pesticide use in a specific region or on a specific farm. Similar simple screening procedures are nonetheless widely applied in pesticide registration throughout Europe (a single scenario is used in the Netherlands and Germany, two scenarios are used in Denmark, and nine scenarios have been defined at the EU level). Since the decision-making process needs to be pragmatic, a more differentiated approach does not seem feasible at EU or national levels. At the farm level, however, the introduction of precision agriculture offers an opportunity to increase control over pesticide leaching. This is especially relevant in regions where ground water is exploited as a resource for drinking water.

In precision agriculture, soil variability is actively managed. Detailed soil information required for this purpose will generally be collected through a specific soil survey (Bouma et al., 1999). When combined with data on pesticide degradation and sorption (either measured or derived from literature), the survey information can be used to model pesticide fate in a large number of soil profiles. Hot spots can be identified and pesticide management can be fine-tuned to avoid excessive leaching to the ground water.

This study evaluates the effects of soil variability and weather conditions on pesticide leaching. The evaluation is conducted at the farm level and aims at identifying management options that will increase control over ground water quality. The study area is located on a 100-ha arable farm in the central-western part of the Netherlands. The farmer is an early adopter of precision agriculture technology and has a soil database containing physical and chemical properties for over 600 soil profiles. Pesticide use, which has been documented since 1995, is evaluated following a step-wise approach. First, a relative risk assessment identifies pesticides that pose a relatively high risk to the environment. The assessment is based on simulations for a representative soil profile and uses literature data on pesticide degradation and sorption. Second, the effect of weather conditions is analyzed through 20 years of simulations for three distinct soil profiles. Results are summarized in cumulative probability plots of simulated leaching. Third, the effects of soil variability are investigated through simulations for all soil profiles included in the soil database. Simulations are conducted for three pesticides presenting the highest environmental risk. Reference conditions are selected on the basis of the cumulative probability plots derived earlier. Fourth and finally, the point data are interpolated and hot spots are identified as areas where the average pesticide concentration in percolating water exceeds the environmental threshold value of 0.1 µg L-1. Possible implications for pesticide management are discussed at the within-field, field, and farm level.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Study Area
Research was conducted on a commercial arable farm in the central-western part of the Netherlands (51°17' N, 4°32' E). The farm covers an area of approximately 100 ha and applies a crop rotation of winter wheat (Triticum aestivum L.), consumption potatoes (Solanum tuberosum L.), and sugar beet (Beta vulgaris L.). Soils originate from marine deposits, are generally calcareous, and have textures ranging from sandy loam to clay. They are characterized as fine, mixed, mesic Typic Fluvaquents (Soil Survey Staff, 1998) or Mn25A-Mn45A on the Dutch 1:50000 soil map (Vos, 1984). Soil variability is large and mainly expressed through differences in texture (the average clay content over 0–100 cm varies from 14 to 50%), soil organic matter (SOM) content (the average SOM content over 0–100 cm varies from 5 to 58 g kg-1), and subsoil composition (peat or mineral matter). Drainage conditions are excellent and controlled through a dense system of pipe drains installed at approximately 1 m depth. In general terms, the area is considered prime agricultural land.

Soil Database
During the spring of 1997, a detailed 1:5000 soil survey was conducted in the study area. Basic soil properties were collected for 612 georeferenced soil-sampling sites and stored in a soil database. Texture and SOM content were estimated in the field using hand texture and color. Estimates were tested against a limited number of laboratory measurements to ensure accurate characterization. Based on this information, soil layers were grouped into classes defined by the Staringreeks (Wösten et al., 1994). This classification distinguishes between topsoil and subsoil layers, which are further differentiated by textural composition and SOM content. Sixteen classes were identified and sampled in the field. The average bulk density and saturated moisture content were determined for each class using at least four replicate samples.

Soil hydraulic characteristics were derived through a continuous pedotransfer function (PTF) developed at the DLO-Staring Center (Wösten et al., 1998). The PTF is based on soil physical measurements for 620 soil samples collected from major soil types in the Netherlands. It relates basic soil properties such as texture, SOM content, and bulk density to a set of Van Genuchten parameters (Van Genuchten, 1980). These parameters describe the moisture retention and hydraulic conductivity curve and were derived for individual soil layers. A sensitivity analysis by Vanclooster et al. (1992) identified saturated moisture content as the most sensitive hydraulic parameter affecting nitrate leaching. Considering their result, measured saturated moisture contents were used to replace PTF estimates.

Simulation Model
Pesticide fate modeling was conducted using the WAVE model (Vanclooster et al., 1994, 1999). The WAVE model had previously been validated on soil moisture and soil mineral N data from the study area. Van Alphen and Stoorvogel (2000) compared simulated soil moisture contents to weekly time domain reflectometry (TDR) measurements for three sites and two depths (20 and 40 cm). The overall coefficient of determination was 66%. Van Alphen (2002) compared simulated and measured soil N data for four soil types and two depths (0–30 and 30–60 cm). The coefficient of determination equaled 84%.

The WAVE model integrates existing models for one-dimensional water flow (SWATRER; Dierckx et al., 1986) and crop growth (SUCROS; Spitters et al., 1988) with specific modules for heat and solute transport from LEACHM (Wagenet and Hutson, 1989). The model is mechanistic–deterministic and solves physical transport equations using a numerical finite difference technique. Soil profiles are described in terms of soil horizons, which are divided into 1-cm compartments. Mass balances are kept for water, heat, and solutes within each compartment, taking into account different sink–source terms.

Water movement is modeled using the Richard's equation (Richard, 1931), which combines the mass balance and Darcian flow equations. Soil hydraulic properties are defined by parametric Van Genuchten equations (Van Genuchten, 1980) and water uptake by crops is modeled after Feddes et al. (1978). Potential uptake is defined by a sink term (d-1) that equals the potential transpiration (mm d-1) divided by the rooting depth (mm). Actual uptake is derived by multiplication with a crop-specific reduction factor (ranging between 0 and 1) that reduces uptake rates at high and low pressure head values.

Mass flux of pesticides is modeled using the convection–dispersion equation. In view of the limited availability of nonlinear sorption parameters, pesticide retention is described by a simple equilibrium sorption isotherm:

[1]
in which X is amount of pesticide sorbed to the solid phase (kg kg-1), Kd is a linear distribution coefficient (L kg-1), and c is the pesticide concentration in the liquid phase (kg L-1). The term Kd is generally derived by multiplying the distribution coefficient over organic matter and water Kom (L kg-1) with the mass fraction of organic matter in the soil. Pesticide degradation is included through a first-order decay model:

[2]
in which R is the rate of degradation (kg L-1 d-1), k is the degradation coefficient (d-1), and c* is the pesticide concentration in the soil system (kg L-1). As proposed by Walker (1974), potential degradation coefficients kpot (d-1) at reference conditions are corrected using:

[3]
in which fT is a factor to correct for the influence of soil temperature and f{theta} is a reduction factor for the effect of soil moisture conditions. The relation between kpot and the half-life time of a pesticide at reference conditions, DT50 (d), is defined by:

[4]
<>Following Boesten (1986), the factor fT is described by:

[5]
in which T is the soil temperature (°C) and Tref is the soil temperature at reference conditions (°C). The factor f{theta} is calculated as:

[6]
in which {theta} is the soil moisture content (g kg-1), {theta}ref is the soil moisture content at reference conditions (g kg-1), and B is a constant.

Inventory and Risk Assessment
Pesticide use in the study area has been documented since 1995. Records contain a listing of chemicals (commercial names) used in the different crops. Active ingredients (chemical names) and recommended application rates were derived from the Dutch crop protection guide (Oomen et al., 1999). This document also provided information on specific restrictions applying to some pesticides (e.g., highly mobile compounds are banned from ground water protection areas).

After the inventory, a relative risk assessment identified pesticides posing a relatively high risk of exceeding the environmental threshold for leaching. The assessment was based on a series of simulations for a representative soil profile reflecting average texture and SOM content in the study area (Fig. 1 , Profile B). Each simulation assumed a standard pesticide application of 1 kg ha-1 on 1 Apr. 1999. Leaching was calculated over a period of one year and expressed as a fraction of the dosage applied. The DT50 and Kom were varied from 1 to 100 (using an interval of 10) with separate runs conducted for each combination. Results were interpolated through ordinary kriging (Journel and Huybregts, 1978) to create a continuous surface of the leaching fraction as a function of DT50 and Kom.



View larger version (35K):
[in this window]
[in a new window]
  		Fig. 1. Sampled soil profiles. Profile A combines heavy texture and high soil organic matter (SOM) percentage (relatively low risk for leaching), Profile B reflects average properties in the study area (average risk for leaching), and Profile C combines light texture and low SOM percentage (relatively high risk for leaching). Sand fractions are not included, but equal 100 minus the sum of clay and silt fractions.

 
Literature values of DT50 and Kom were available for all pesticides used in the study area (Linders et al., 1994). As far as both properties were <=100, existing combinations were plotted on the leaching surface. In this way a leaching fraction could be estimated for each compound. Estimates were subsequently combined with recommended application rates to identify pesticides carrying a relatively high risk of exceeding a concentration of 0.1 µg L-1 in percolating water. Strongly sorbing compounds (Kom > 100 L kg-1) were not considered as they present little environmental risk (two compounds with a DT50 > 100 d were included in this category).

Laboratory and Field Measurements
A series of laboratory and field measurements was conducted to produce validation data for the WAVE model. These data were required to complete the model's validation status, which so far lacked a specific performance analysis for pesticide fate. Measurements were limited to a single pesticide for financial reasons. Isoproturon, a common herbicide used in winter wheat, was selected for its frequent use in practice and the availability of relatively low-cost measurement kits. EnviroGard Isoproturon Plate Kits (Strategic Diagnostics, Newark, DE) provide a quantitative laboratory test for the detection of isoproturon residues in soil moisture samples. They resemble ELISA antibody tests and can measure concentrations in the range of 0.05 to 0.5 µg L-1.

Field measurements were conducted on a single 15.7-ha winter wheat field. Soil moisture samples were taken before isoproturon application on 20 March and again at 4 and 26 wk after application. Ten sampling sites were positioned strategically across the field to represent textural and SOM variation. Samples were collected using porous cups (10-cm length and 2-mm diameter) installed at 20- and 40-cm depths. Three sites were also used for ground water sampling, which was conducted once at 26 wk after application.

Laboratory measurements determined the DT50 and Kom of isoproturon. Measurements were conducted in undisturbed 300-cm3 soil cores collected from three sites at 20- and 40-cm depths (three replicates, 18 samples in total). Sampled soil profiles were selected to reflect textural and SOM variation in the study area (Fig. 1). To determine DT50, 1 µg of isoproturon was applied to each soil core. Samples were incubated at 18°C and moisture contents were maintained by regular additions of distilled water as necessary. At 1, 14, and 28 d the samples were analyzed for isoproturon residues. Measurements were conducted in suspensions of 5 g fresh soil in 25 mL distilled water. The shaking period was set at 24 h. Gravitational soil moisture contents were determined for each sample through overnight drying in an oven at 110°C. The DT50 was calculated through linear regression of log-transformed concentrations (corrected for soil moisture contents) against time.

Sorption characteristics were measured in disturbed samples collected together with the soil cores for DT50 (three sites, two depths). Organic C fractions were determined for each sample using an Interscience (Ontario, Canada) Elemental Analyzer EA1108. Conversion from total C to SOM was made assuming a factor 2.0 for the ratio of SOM to total C (Nelson and Sommers, 1982). Next, 5 g of air-dry soil from each sample was added to 25 mL of a 1, 5, and 10 µg L-1 isoproturon solution. The resulting 18 suspensions (six samples, three concentrations) were shaken for a period of 24 h. The Kom (L kg-1) was calculated as:

[7]
in which Io is the original isoproturon concentration (kg L-1), Ir is the isoproturon concentration recovered after shaking (kg L-1), W is the volume of water (L), S is the amount of soil material (kg), and SOM is the soil organic matter fraction (kg kg-1).

Modeling Pesticide Fate
Effect of Weather Conditions
Apart from chemical and soil properties, the extent to which pesticides leach is largely dependent on weather conditions. Rainfall and evapotranspiration vary greatly between years, resulting in different quantities of percolating water and solute transport. This effect was investigated through 20 years (1980–1999) of simulations for Soil Profiles A, B, and C in Fig. 1. Historical weather data (rainfall, temperature, and potential evapotranspiration) were available from a regional weather station situated 20 km from the study area. Pesticide application (1 kg ha-1 on 1 April) and chemical properties (DT50 and Kom) were identical in all simulations. The latter were specified using average values measured for isoproturon. Leaching was calculated over a period of one year and expressed as a fraction of the dosage applied.

After completing the simulations, leaching fractions from each soil profile were summarized in cumulative probability plots. The shape of these plots was compared to test the hypothesis that though leaching may differ in absolute terms, the probability increment is independent of soil type. In other words, weather conditions and soil variability act independently on pesticide leaching. Once this hypothesis was confirmed, the year that consistently matched 90% probability was identified (higher leaching occurred in 2 out of 20 years). This year provided the reference conditions for a final series of simulations aimed at evaluating the effects of soil variability.

Effect of Soil Variability
Effects of soil variability on pesticide leaching were analyzed at the within-field, field, and farm level. All profiles in the soil database were included in a simulation series to identify hot spots, areas of land where the average pesticide concentration in percolating water exceeds 0.1 µg L-1. Within the calculation we assume that the ground water bodies function as a buffer and that the concentrations in ground water correspond with the average concentration in percolating water. Pesticide leaching was calculated at the point level as the amount of chemical (kg) transported below 1 m depth divided by the total downward flux of soil water (L). Point data were subsequently interpolated using ordinary kriging.

Simulations were conducted for a limited number of compounds identified in the relative risk assessment. Pesticide application was specified after recommendations in the Dutch crop protection guide (Oomen et al., 1999). Meteorological reference conditions were provided by the year matching 90% probability in the probability plots. This introduced a 10% risk level: soils showing up with no or acceptable leaching may exceed environmental thresholds at a maximum of 1 out of 10 years.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Inventory and Risk Assessment
Table 1 presents an overview of pesticides used in the study area during the period 1995–1999. The DT50 and Kom were derived from Linders et al. (1994), who compiled average values based on a review of available laboratory measurements.


View this table:
[in this window]
[in a new window]
  		Table 1. Pesticides used in the study area during the period 1995–1999. Half-lives (DT50) and sorption coefficients (Kom) were derived from Linders et al. (1994). Recommended application rates are specified after Oomen et al. (1999).

 
Figure 2 presents simulated leaching from a representative soil profile (Fig. 1, Profile B) as a function of DT50 and Kom. Used pesticides are represented by their DT50–Kom combinations, which are plotted on top of the interpolated leaching surface. By combining the two sources of information, a leaching fraction was estimated for each compound. Estimates were subsequently multiplied by a recommended application rate and divided by the simulated percolation at 1 m depth (531 mm). Resulting average concentrations identified isoproturon, metribuzin, and bentazon as pesticides presenting a risk of exceeding the threshold concentration (0.1 µg L-1) for percolating water. The risk associated with these pesticides differed, with isoproturon presenting the lowest risk and bentazon presenting the highest risk.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 2. Leaching (in percent) as a function of degradation time (DT50) and coefficient of sorption to organic matter (Kom). The DT50–Kom combinations of pesticides used in the study area are plotted as {diamond} (leaching <=0.1 µg L-1) or {diamondsuit} (leaching >0.1 µg L-1).

 
Measurements and Model Validation
Isoproturon, one of the three highest-risk pesticides, was selected for field and laboratory measurements. Sorption characteristics measured in disturbed soil samples from Profiles A, B, and C (Fig. 1) resulted in an average Kom of 63 (±16) L kg-1. Compared with the average literature value in Table 1 this meant an increase of 9 L kg-1 or 17%. Sorption did not differ significantly between concentrations considered (1, 5, and 10 µg L-1). Degradation patterns recorded in incubated soil cores are presented in Fig. 3 . The data are plotted as residual concentrations (in percent) relative to the amount recovered after 1 d of incubation. Degradation followed first-order reaction kinetics with an exponential decrease over time. Linear regression of log-transformed concentrations against time of incubation resulted in an average DT50 of 19.4 (±8.9) d. This meant a decrease of 10.6 d or 35% compared with the average literature value in Table 1. The average gravitational soil moisture content of incubated soil cores was 20 (±2)%. Degradation did not differ significantly between topsoil and subsoil samples.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 3. Degradation patterns of isoproturon recorded in incubated soil cores. Samples are coded according to their soil profile (A, B, or C) and sampling depth (1 for topsoil, 2 for subsoil). Sample B1 is excluded as pesticide concentrations failed to show a consistent decrease over time.

 
Isoproturon residues measured in the field were used to validate the WAVE model. Moisture samples collected just before isoproturon application on 20 March showed no signs of an initial concentration present in the soil. Samples collected 4 wk later detected isoproturon residues at 20 cm depth (3.13 [±1.14] µg L-1), while no residues were found at 40 cm depth (<0.05 µg L-1). After 28 wk, residues were found at both depths: 0.46 (±0.15) µg L-1 at 20 cm and 0.29 (±0.03) µg L-1 at 40 cm. No residues were detected in ground water samples at this time (<0.05 µg L-1).

Field measurements were compared with simulated concentrations for a representative soil profile (Fig. 1, Profile B). Model input was specified using a combination of measured chemical properties (DT50 and Kom) and standard parameter settings described by Boesten and Van der Linden (1991). Meteorological data were available from an on-farm weather station. Simulation started on 1 Jan. 2000 with no isoproturon residues in the soil (measured initial condition). Four weeks after application (18 April), the model calculated concentrations of 3.04 µg L-1 (20 cm) and 0.0 µg L-1 (40 cm). This was well within one standard deviation of measured data. Simulated and measured concentrations at 28 wk after application (5 October) are presented in Fig. 4 . Under the assumption that the entire isoproturon dosage (1 kg ha-1) entered the soil environment, the simulation overestimated measured concentrations by approximately a factor of 2. This was not surprising since losses during spraying (e.g., spray drift, interception, volatilization, photochemical degradation) were not considered. For illustration purposes, the simulated concentrations were corrected for crop interception and added to Fig. 4. Correction was performed using a simple leaf area index based factor (exp[-0.5 LAI]) derived for cereals by Gyldenkærne et al. (1999). At the time of application, the simulated LAI equaled 0.8 m2 m-2. The corrected application rate (0.67 kg ha-1) reduced the simulated concentrations to a level within one standard deviation of the measured data (verified at 4 and 28 wk after application).



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 4. Simulated and measured isoproturon concentrations at 28 wk after application. Simulated concentrations for the higher dosage (1 kg ha-1) assume no initial losses, whereas the lower dosage (0.67 kg ha-1) accounts for interception losses during spraying.

 
Modeling Pesticide Fate
Cumulative probability plots of pesticide leaching from Soil Profiles A, B, and C (Fig. 1) are presented in Fig. 5 . Each plot is based on 20 years of simulations (1980–1999) using isoproturon as a reference chemical. Though absolute values differed, the shape of the probability plots was similar among the soils. This confirmed the hypothesis that weather conditions and soil properties act independently on pesticide leaching. The year that consistently matched 90% probability was 1981.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 5. Cumulative probability plots of isoproturon leaching from soil profiles (A, B, and C). Plotted data refer to simulated leaching fractions for 20 years of data.

 
Effects of soil variability on pesticide leaching are reflected in Fig. 6 . Presented patterns were derived for risk-bearing chemicals (isoproturon, metribuzin, and bentazon) using 1981 as a reference for simulation. The latter implies that calculated pesticide concentrations in percolating water may be exceeded in 1 out of 10 years (10% risk level). Pesticide application was specified after Oomen et al. (1999): 1 kg ha-1 of isoproturon on 1 April, 0.1 kg ha-1 of metribuzin on 1, 15, and 30 May (0.3 kg ha-1 in total), and 0.48 kg ha-1 of bentazon on 15 and 30 May (0.96 kg ha-1 in total). As initial losses could not be quantified unambiguously for all three chemicals, the entire dosage was assumed to enter the soil environment. The DT50 and Kom were measured (isoproturon) or specified after literature values (metribuzin and bentazon). At the farm level, simulated average concentrations in percolating water equaled 0.07 (±0.07) µg L-1 for isoproturon, 0.20 (±0.11) µg L-1 for metribuzin, and 6.76 (±0.96) µg L-1 for bentazon. Simulated leaching was considered no risk if average pesticide concentrations remained <=0.1 µg L-1. Concentrations within the range of 0.1 to 0.2 µg L-1 were considered low risk, as initial losses were neglected and validation had indicated an overestimation of measured concentrations. Average pesticide concentrations > 0.2 µg L-1 were considered high risk.



View larger version (65K):
[in this window]
[in a new window]
  		Fig. 6. Effects of soil variability on the leaching of isoproturon, metribuzin, and bentazon.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
The relative risk assessment, which identified isoproturon, metribuzin, and bentazon as higher-risk pesticides, was based on simulated leaching under chosen reference conditions. Soil properties and application dates were fixed. Differences among pesticides originated from chemical properties (DT50 and Kom) and application rates, not from different periods of application. This pragmatic simplification was required to limit the number of simulations. Its effect is mitigated by the fact that autumn application, which presents a higher risk for leaching, is not practiced in the study area.

Specific attention was required for the validation status of the WAVE model. Results showed that simulated isoproturon concentrations were accurate 4 wk after application. This was mainly interpreted as a confirmation that downward progression of the pesticide front was accurately modeled. At 28 wk, however, measured concentrations were overestimated by approximately a factor of 2. Performance improved after correction for crop interception, allowing discrepancies to be largely attributed to initial losses. Model performance within the soil environment was considered accurate.

The relevance of initial losses is undisputed, especially for pesticides that are applied after crop emergence. However, quantifying these losses for different pesticides and crops is difficult with the current knowledge (Boesten, 1999). In fact, the pragmatic choice to neglect the processes involved (spray drift, interception, surface volatilization, and photochemical degradation) may be one of the main reasons that pesticide degradation is generally underestimated by simulation models (Beulke et al., 2000). The correction factor for crop interception, which was used in the validation experiment for isoproturon, is specific to cereals. As metribuzin and bentazon are applied to potato, this factor could not be applied unambiguously. Moreover, crop interception is merely one of several processes causing initial losses. An alternative nonmechanistic way of dealing with initial losses was applied in interpreting simulated leaching patterns (Fig. 6). Apart from no-risk and high-risk categories, a low-risk category accounted for concentrations that exceeded the environmental threshold by less than a factor of 2. This factor corresponded to the approximate overestimation found during model validation.

The main consideration in analyzing the effect of weather conditions was to select a reference set that represented severe, but not extreme leaching. To minimize subjectivity, the selection process was based on cumulative probability plots (Fig. 5). These clearly showed that leaching was extreme in 2 out of 20 years (notice the abrupt increase of the leaching fraction between 90 and 95% probability). The year representing severe but non-extreme leaching was selected as the year matching 90% probability. The fact that this introduced a 10% risk level was considered acceptable.

Effects of soil variability on pesticide leaching were found to be significant, though variable among compounds. Bentazon clearly presented the highest environmental risk. Simulated average concentrations in percolating water consistently exceeded 0.1 µg L-1 and did so at an impressive margin. Due to its low adsorption coefficient (0.4 L kg-1), spatial variation of simulated leaching was low (the coefficient of variation at the farm level equaled 0.14). Isoproturon presented a very different situation. Spatial variation of simulated leaching was high (coefficient of variation at the farm level equaled 1.00), but concentrations in percolating water only exceeded 0.1 µg L-1 at the within-field level. High-risk areas (or hot spots) were small, well confined and limited to a few fields. Metribuzin presented the most interesting case. Spatial variation of simulated leaching was substantial (coefficient of variation at the farm level equaled 0.55) and concentrations in percolating water exceeded 0.1 µg L-1 at the within-field, field, and farm levels. In terms of environmental risk, differences within and among fields were apparent. High-risk areas were found in 9 out of 10 fields. Compared with isoproturon these areas were larger, though still well confined (not scattered). Four fields presented average concentrations <=0.2 µg L-1 (no risk or low risk) and six fields exceeded 0.2 µg L-1 (high risk). The average metribuzin concentration at the farm level equaled the maximum for the low-risk category (0.20 µg L-1).

A key question is how simulated leaching patterns can be used to facilitate precision management of pesticides. Automatically, this raises the issue of scale: at which level (within-field, field, or farm level) should the environmental threshold concentration be implemented? Obviously, each scale level has different implications for pesticide management. Implementation at the within-field level requires pesticide concentrations to remain <=0.1 µg L-1 everywhere. In the current context this would restrict pesticide use on all fields containing hot spots, irrespective of average field concentrations. Especially in the case of isoproturon this would exaggerate the environmental risk. Implementation at the farm level focuses on average pesticide concentrations calculated for the entire farm. Results would in this case support the use of isoproturon (no risk) and metribuzin (low risk). Bentazon (high risk) would be banned (in fact this chemical is scheduled for removal from the Dutch market by 2004). One can argue that the generalization inherent to this scale level is undesirable, especially for large farms. Moreover, farm-wide application during a single growing season is unrealistic for most pesticides. A third and intermediate option is implementation at the field level. It seems the most realistic option, as management operations are generally planned and executed at this level. In the current context a field-level evaluation would support the use of isoproturon, impose a ban on bentazon, and restrict the use of metribuzin. With respect to the latter, several options may be considered. The simplest option would be to allow metribuzin to be used only on no-risk and low-risk fields. High-risk fields (in this case 6 out of 10) should be treated with an alternative, less-persistent chemical. A more sophisticated approach might involve site-specific application of metribuzin to no-risk and low-risk areas, combined with application of an alternative chemical to high-risk areas. Though technically possible, the simple option seems more feasible.

As a final remark, it should be mentioned that results presented for isoproturon are based on measured chemical properties, whereas results for metribuzin and bentazon are based on average values found in literature. Laboratory measurements for isoproturon differed from literature values by -35% (DT50) and +17% (Kom). Various sensitivity analyses of pesticide leaching models have indicated that these parameters (especially DT50) are important (e.g., Jury et al., 1987; Boesten and Van der Linden, 1991). Ideally, chemical properties of all risk-bearing pesticides would have been measured in local soil samples. At present this is very costly and proved unfeasible within this research program. The next best option was to use available measurements and complement these with literature values. Obviously this introduced different levels of uncertainty to be associated with model output. In the future, the increased availability of low-cost measurement kits (as used in this study) can help to overcome this limitation.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
(i) Out of a total of 19 pesticides, the relative risk assessment identified isoproturon, metribuzin, and bentazon as relatively high risk for leaching. Risk levels differed strongly, with isoproturon presenting the lowest risk and bentazon presenting the highest risk.

(ii) Comparison of cumulative probability plots for different soil profiles confirmed that weather conditions and soil properties act independently on pesticide leaching. Probabilistic characterization of historical weather data can thus be used to select reference conditions for pesticide fate modeling.

(iii) Soil variability proved an influential factor affecting pesticide leaching at the within-field, field, and farm levels. Spatial variability was highest for isoproturon, followed by metribuzin and bentazon. Differences were related to sorption characteristics: stronger sorption resulted in higher variation.

(iv) The presented step-wise evaluation can serve as a basis for identification and precision management of high-risk pesticides. Management implications largely depend on the scale level (within-field, field, or farm level) at which the environmental threshold concentration of 0.1 µg L-1 is implemented. A political decision regarding this issue is therefore required.


   ACKNOWLEDGMENTS
 
The research of Jetse Stoorvogel is funded by a fellowship of the Royal Netherlands Academy of Arts and Sciences. We wish to thank Jos Boesten for helpful discussions and comments on the manuscript. Piet Peters provided valuable assistance during the field experiments and the Van Bergeijk family is acknowledged for supporting this research on their farm.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow 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 ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by van Alphen, B. J.
Right arrow Articles by Stoorvogel, J. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by van Alphen, B. J.
Right arrow Articles by Stoorvogel, J. J.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by van Alphen, B. J.
Right arrow Articles by Stoorvogel, J. J.
Related Collections
Right arrow Spatial Distribution
Right arrow Ground Water Quality
Right arrow Best Management Practices
Right arrow Agricultural Pesticides
Right arrow Water Pollution

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