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)
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 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 Google Scholar
Google Scholar
Right arrow Articles by Wolters, A.
Right arrow Articles by Vereecken, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wolters, A.
Right arrow Articles by Vereecken, H.
Agricola
Right arrow Articles by Wolters, A.
Right arrow Articles by Vereecken, H.
Related Collections
Right arrow Pesticides
Right arrow Organic Compounds
Right arrow Other Models
Right arrow Agricultural Pesticides
Right arrow Air Pollution
Published in J. Environ. Qual. 33:1629-1637 (2004).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Atmospheric Pollutants and Trace Gases

An Improved Description of Pesticide Volatilization

Refinement of the Pesticide Leaching Model (PELMO)

André Woltersa,*, Michael Kleinb and Harry Vereeckena

a Forschungszentrum Jülich GmbH, Institute of Chemistry and Dynamics of the Geosphere IV: Agrosphere, 52425 Jülich, Germany
b Fraunhofer-Institute for Molecular Biology and Applied Ecology, P.O. Box 1260, 57377 Schmallenberg, Germany

* Corresponding author (a.wolters{at}fz-juelich.de).

Received for publication October 17, 2003.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The consideration of pesticide volatilization from soil surfaces as an integral component of pesticide fate models is of importance, especially as part of the Predicted Environmental Concentrations (PEC) models used in the registration procedures for pesticides. The Pesticide Leaching Model (PELMO), which is used in the European registration process, was modified to allow for a reliable prediction of volatilization from soil. The previous PELMO version was upgraded by improving the spatiotemporal discretization at the soil surface, improving the empirical description of temperature dependence of Henry's law constants and including increased sorption of pesticides in dry soils. Comparison of predictions with experimental findings revealed the improvements of PELMO to contribute to a more realistic reflection of measurements, particularly at initial stages of the studies. The broad range of literature values of Henry's law constants was shown to have a significant effect on predicted volatilization fluxes. As a main refinement, the tendency of pesticides toward enhanced volatilization under moist conditions was correctly calculated by the improved model. Variations between model predictions and measurements were due to a lack of experimental data on soil sorption under dry conditions and indicated the need for further calibration of the model. The description of water content in the top layer was subject to uncertainty, which was exemplified by an overestimation of soil moisture during the last days of the field study. Thus, future model improvement will be dependent on experimental support to obtain more detailed information on soil–air–water partitioning of pesticides in the top soil layer.

Abbreviations: FOCUS, Forum for the Co-Ordination of Pesticide Fate Models and Their Use • PEARL, Pesticide Emission Assessment at Regional and Local Scales • PEC, Predicted Environmental Concentrations • PELMO, Pesticide Leaching Model


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
PESTICIDES ARE APPLIED with the intent of maximizing efficacy while minimizing off-site movement. However, their judicious use demands a practical knowledge of their fate and effect on agricultural and natural ecosystems (Hapeman et al., 2003). For most pesticides, volatilization is as important as chemical and microbiological degradation in causing the dissipation of pesticides from soil. Volatilization decreases the amount of a pesticide available for control of pests and the potential for ground water contamination but increases the potential for contaminating the atmosphere and surface water. Since many pesticides are considered to be toxic or carcinogenic (Doull, 1989), volatilization may increase the risk of illness to persons living downwind from treated fields (Yates, 1993).

Direct determination of pesticide volatilization in the field can be achieved by using micrometeorological methods (Taylor and Spencer, 1990; Woodrow et al., 1990), such as the aerodynamic gradient and the Bowen ratio methods. Beside various climatic parameters, the concentration of active ingredients in the air is to be measured at different heights above the field. Substance flux in the atmosphere can be calculated using various mathematical models (Majewski, 1999). Alternatively, wind-tunnel systems have been developed to approximate field conditions as closely as possible (e.g., Stork et al., 1994; Bedos et al., 2002). The wind tunnel used for the studies performed within this work allows for direct measurement of volatilization and biomineralization under field-like conditions, in combination with the advantages of laboratory facilities (e.g., use of radioisotopes) (Stork, 1995).

Volatilization from soil is a complex system, requiring the balancing of several processes. Pesticides in soil will partition between soil solids, interstitial soil solution, and gas-filled soil pores (Cousins et al., 1999). For a sorbed chemical to volatilize from the surface of the soil, it must first desorb from the soil solids into the soil interstitial solution, from where it can partition into the soil air. Once in the soil air at the surface, there is the potential for transportation across the boundary layer and into the bulk atmosphere. Transfer across this boundary layer is via diffusion. The magnitude of the diffusive flux will be determined by the concentration gradient between the atmospheric and the soil air compartments. It will also be affected by environmental variables such as temperature, surface roughness, and soil characteristics. The situation is made more complex in that some of these variables are interlinked. For example, temperature will affect the effective diffusion coefficient in soil by its influence on the free air diffusion coefficient and also the soil moisture content, which will affect the partitioning of the pesticides between the soil solids and the air phase. These interdependencies make the prediction of pesticide volatilization by means of mathematical models more difficult. For an environmental risk assessment it is important that these interaction effects can be embedded in model approaches. During the last decades several approaches for the prediction of pesticide volatilization have been developed, which reflect the crucial soil process with varying degree of accuracy (e.g., Jury et al., 1983; Baker et al., 1996; Wang et al., 1997; Woodrow and Seiber, 1997; Reichman et al., 2000; Chen and Rolston, 2000), covering the range from empirical models to more sophisticated models.

The simulation of volatilization behavior as an integral component of a complete pesticide transport model is of utmost importance. Attention should be paid to the fact that many PEC (Predicted Environmental Concentrations) models used for pesticide registration, including those used by the USEPA, do not take volatilization losses into account when evaluating surface water and ground water contamination risks (Vanclooster et al., 2000). In a previous work, model predictions using the European registration models PELMO (Pesticide Leaching Model; Klein, 1995) and PEARL (Pesticide Emission Assessment at Regional and Local Scales; Leistra et al., 2001; Tiktak et al., 2000) were compared with experimentally determined volatilization rates and revealed limitations of the current models, such as the uppermost compartment thickness, making an enormous influence on predicted volatilization losses (Wolters et al., 2003). Furthermore, both models did not reflect the soil moisture dependence of sorption coefficients and markedly underestimated volatilization at the initial stage of the studies, illustrating the need for the development of an advanced model to remove these restrictions (Van den Berg et al., 1999; Wolters et al., 2002; Ferrari et al., 2003).

This paper is a follow-up to the publication by Wolters et al. (2003) and summarizes the improvements and refinements of the PEC model PELMO with regard to the description of volatilization from bare soil. Model predictions were evaluated on the basis of the results of a wind-tunnel study, which has already been used for the evaluation of the previous PELMO version. For the verification of the findings and for establishing the credibility of the model results, calculations were compared with volatilization rates of pesticides measured in the field by micrometeorological methods.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Wind Tunnel: Experimental Setup and Analyses
For an evaluation of the model approach, volatilization rates determined in a wind-tunnel study of 13 d after simultaneous application of four pesticides, covering a broad range of physicochemical properties (Table 1), were compared with the model output. Carbon-14-labeled pesticides parathion-methyl [O,O-dimethyl O-(4-nitrophenyl) phosphorothioate], fenpropimorph [cis-4-[3-[4-(1,1-dimethylethyl)phenyl]-2-methylpropyl]-2,6-dimethylmorpholine], and terbuthylazine [6-chloro-N-(1,1-dimethylethyl)-N'-ethyl-1,3,5-triazine-2,4-diamine] and nonlabeled chlorpyrifos [O,O-diethyl O-(3,5,6-trichloro-2-pyridinyl) phosphorothioate] were applied to gleyic cambisol in accordance with agricultural practice by using a semiautomatic sprayer (Stork et al., 1994). A detailed description of the measurement and monitoring device, the analytical procedure, the experimental soil, and an explicit illustration of experimental findings can be taken from Wolters et al. (2003). Reliability and repeatability of the used wind-tunnel system have been verified previously (Stork, 1995).


View this table:
[in this window]
[in a new window]
  		Table 1. Physicochemical active ingredient data of the investigated compounds (Tomlin, 2000) used in the Pesticide Leaching Model (PELMO).

 
Field Study: Field Site and Application
An evaluation of the model predictions under field conditions was performed on the basis of a study performed by Stork et al. (1996)( 1998). Experimental soil was classified as an orthic luvisol, and the Ap horizon (plow layer) contained 1.1% organic carbon, 6.4% sand, 78.2% silt, and 15.4% clay. The bulk density of the Ap horizon was about 1.57 g cm–3 and wilting point was set to 14.9%vol. Before starting the experiment, the background concentration of the test compounds was determined in the soil analyses from samples collected to a depth of 5 cm. Parathion-methyl (wettable powder, 40% a.i., 604 g a.i. ha–1), fenpropimorph (emulsified concentrate, 75% a.i., 1440 g a.i. ha–1), and terbuthylazine (suspension concentrate, 50% a.i., 1180 g a.i. ha–1) had been applied as a spray mixture (applied water: 202 L ha–1) simultaneously by a field sprayer. The micrometeorological methods used for volatilization measurement included aerodynamic and Bowen-ratio methods (Majewski et al., 1989, 1990). Various sensors were used to gauge the meteorological data, including wind speed, wind direction, water vapor pressure, air temperature, soil temperature, soil heat flux, net radiation, and soil surface moisture. At 13 points in time over the 20-d field study concentrations of pesticides in air were measured for calculation of the volatilization fluxes by the two micrometeorological methods. Volatile compounds were collected during sampling intervals of 1 to 2 h at three heights (0.2, 0.5, and 0.8 m) by sampling air at a flow rate of 50 L min–1 using a glass tube (35-mm i.d.) packed with an adsorbent (10 g of XAD-4; Amberlite Sigma, Deisenhofen, Germany). Care was taken to ensure that the upwind fetch was always more than 100 times larger than the distance of the highest air sampling point from the soil surface. The air-sampling rates were measured with a flow meter and a gas meter was used for measuring the total air volumes of each sample. For each sampling, the upwind background concentrations of the pesticides were determined at a height of 1.5 m. A comprehensive illustration of the procedure and the experimental findings is given by Stork et al. (1996)(1998).

PELMO: Improved Description of Volatilization from Soil
The official version of PELMO used for registration purposes (FOCUS [Forum for the Co-Ordination of Pesticide Fate Models and Their Use] PELMO) estimates volatilization from soil using a simple volatilization module based on Fick's and Henry's law (Klein, 1995). It is assumed that the concentration of the pesticide in the air above the soil is negligibly low. The following equation is used for the calculation of the volatilization rates:

[1]
where J is the volatilization rate (g cm–2 d–1), D is the diffusion coefficient in air (cm2 d–1), H' is the nondimensional Henry's law constant, d is the air boundary layer (cm), and csol is the pesticide concentration in the soil water (g cm–3).

The standard scenario for PELMO simulations implies default values for soil layer thickness (5 cm) and volatilization depth (1 mm; thickness of the soil layer actively involved in the volatilization process). A Freundlich-type equation is included for the consideration of pesticide sorption.

The official PELMO version was enhanced and the following improvements of the volatilization module were implemented.

Based on Henry's law constants measured or estimated (doubling of Henry's law constant every 10°C) at two temperatures, exponential approaches were included for calculating the temperature dependence of water–air partitioning over the relevant temperature range, enabling PELMO to assess Henry's law constants over the course of the study for actual soil temperatures.

A small soil compartment of 1 mm was added on top of the soil for a more realistic reflection of the pesticide concentration at the surface layer, especially immediately after spraying. After pesticide application, all of the applied mass is assumed to be deposited into this surface layer. To simulate the drying out at the soil surface, the minimum soil moisture for that particular surface soil compartment was set to a water content corresponding to air-dried soil. All other soil characteristics were the same as for the first regular soil compartment.

The water-balance equation for the surface soil layer is principally based on the soil moisture of the previous day, precipitation, snowmelt, runoff, percolation, and evapotranspiration. Runoff is estimated based on the runoff-curve-number approach; percolation occurs if soil moisture is above field capacity and excess water is rapidly drained from the soil compartment to deeper soil layers within one time step. The actual evapotranspiration taken up from the different soil compartments is weighted using a triangular function (maximum weighing factor at the soil surface). A more detailed description is given in Klein (1995). Due to its small size of 1 mm, the lower top compartment size results in changing soil moisture of the top millimeter over the course of the study (Fig. 1A and 5).



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 1. Modifications included in the improved PELMO version. (A) Measured and calculated soil moisture profiles over the course of the wind-tunnel study on gleyic cambisol using the official PELMO version (FOCUS PELMO, top compartment thickness: 5 cm) and the modified version (top compartment thickness: 1 mm). (B) Correction factors for soil adsorption coefficients (gleyic cambisol). For moistures in the top millimeter of the soil ranging from above air-dry conditions to below wilting point, the sorption coefficients were increased by the given correction factors.

 
The moisture dependence of the soil adsorption coefficients at low water content (e.g., Petersen et al., 1995) was taken into consideration. For moisture contents ranging between air-dry conditions and wilting point, an increase of the soil sorption coefficient was assumed to occur in the top millimeter of the soil, as exemplified for gleyic cambisol in Fig. 1B. The sorption coefficient in air-dry soil, whose water content was estimated to be 10% of the water content at the wilting point, was increased by a factor (Fss) of 100 independent of the simulation and the pesticide. The dependence of the sorption coefficient on moisture content between air-dry conditions and wilting point is described using a power function:

[2]

[3]
where Fss is the increase of soil sorption when soil is air dry (unitless), {Theta} is the volume fraction liquid phase (m3 m–3), {Theta}wp is the volume fraction liquid phase at wilting point (m3 m–3), and {Theta}ad is the volume fraction liquid phase when soil is air dry (m3 m–3). Above the wilting point, the sorption coefficient remained unchanged. The exponent m is given by:

[4]

Optionally, the improved version enabled volatilization fluxes to be calculated in an hourly resolution, subject to the condition that environmental data were also provided on an hourly basis. For calculations on an hourly basis, an additional procedure was implemented in PELMO to consider non-equilibrium sorption immediately after application. For each compound, a linear increase of sorption coefficients during the first 3 h after application to the final values given in Table 1 was calculated. The physicochemical inputs of the applied compounds required for PELMO calculations are summarized in Table 1.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Wind-Tunnel Study: Volatilization Rates Calculated by the Improved PELMO Version
For evaluation of the volatilization description included in PELMO, volatilization fluxes determined in a wind-tunnel study after simultaneous application of parathion-methyl, fenpropimorph, terbuthylazine, and chlorpyrifos to gleyic cambisol were compared with predictions of the previous and improved versions of PELMO (Fig. 2). The main objective of this comparison was to check whether or not the new modules in PELMO result in a better description of volatilization; thus, the focus in this study was not a perfect mimicry of the experimental data based on optimized calibration of various input parameters.



View larger version (39K):
[in this window]
[in a new window]
  		Fig. 2. Measured (wind-tunnel study) and predicted (PELMO) volatilization fluxes after soil surface application to gleyic cambisol (semilogarithmic plots). (A) Parathion-methyl. (B) Fenpropimorph. (C) Terbuthylazine. (D) Chlorpyrifos.

 
The "previous version" is based on the official FOCUS PELMO version used for registration proposes in Europe. An exponential approach was already included for calculating the temperature dependence of Henry's law constants. As Henry's law constant (taken from literature, no calibration) and the boundary air layer above the first soil compartment (1 mm, no calibration) were the driving input parameters for the estimation of volatilization rates in the official FOCUS PELMO, the calculations were obtained without any calibration steps. A detailed comparison of calculations using FOCUS PELMO with experimental findings is given in Wolters et al. (2003). The improved PELMO version requires additional input parameters, such as the new parameter Fss (Eq. [2]). The linear increase of sorption during the first hours after application was set in the PELMO code and thus could not be used for calibration purposes.

Volatilization rates predicted by the improved PELMO model are markedly higher than those calculated by the previous version; for fenpropimorph and terbuthylazine the computed values even exceeded the measured values at the initial stage of the study by a factor of 20 (Fig. 2B and 2C). Despite a marked increase of the predicted volatilization rates of parathion-methyl, its volatilization at the early stage of the study was still underestimated (Fig. 2A). Chlorpyrifos revealed the best agreement between measured and calculated values on the first day (Fig. 2D). Calculated fluxes of all compounds decreased by Day 8 and increased enormously when irrigation was given, indicating that the improved PELMO version obviously mirrors the tendency of pesticides toward enhanced volatilization under moist conditions (Spencer et al., 1973). This is a main advance in comparison with those PEC models currently in use in European registration procedures; for example, both the previous PELMO version and the PEARL model calculate decreasing volatilization fluxes with increasing water content (Wolters et al., 2003). However, the increase of volatilization predicted by PELMO was much higher than experimentally determined and during the following days predictions exceeded measurements by up to an order of magnitude.

The strong increase of volatilization fluxes in comparison with the calculations of the previous PELMO version is attributed to the reduction of the top compartment size from 5 cm to 1 mm leading to the calculation of higher pesticide concentrations at the soil surface and subsequently causing higher volatilization predictions, as can be taken from Eq. [1]. Drying out and remoistening of the top soil layer were calculated to occur much faster after reducing the compartment size, for example, leading to a large rise in calculated soil moisture from 0.5 to 34%vol within 24 h after irrigation was given (Fig. 1A). In combination with the correction factor included for describing the increase of soil sorption at low water contents (Fig. 1B), this effect enhances the effect on predicted volatilization rates, consequently resulting in an overestimation of the increase of volatilization caused by irrigation. To prevent an overestimation of the soil moisture effect, a reliable estimation of the correction factor is required. Further progress might be achieved considering experimentally determined soil–water partitioning coefficients under low-water-content conditions. Instead of using an exponential approach resulting in a uniform correction factor for all applied compounds to estimate the effect of decreasing water content (Fig. 1B), experimental values would allow for a more detailed reflection of different sorption tendencies of pesticides. For this purpose, a completely new experimental setup to determine soil–air–water partitioning and its dependence on temperature and soil moisture was constructed (German Patent Application no. 101 62 852.8). As the use of a correction factor in the range between wilting point and air-dry conditions is subject to considerable uncertainty, the application of this setup will also allow for the quantification of the effect of the soil type on sorption processes. The development of suitable approaches for calculation of soil–air partitioning might also be considered in future model concepts. In previous studies, equations for calculating organic vapor partition coefficients in unsaturated soils as a function of soil–water content were proposed (Chen et al., 2000). As such calculations also require reliable measurements, such as water vapor sorption isotherms and Henry's law constants, experimental support appears to be indispensably necessary for enhanced modeling.

In addition, calibration of the volatilization approach is required for adjusting predictions to experimental findings as closely as possible, especially with regard to the major impact of the top compartment size of 1 mm on model predictions. For an advanced understanding of the processes occurring at the soil surface, measurements of water content and pesticide concentrations in the top soil layer are required (see discussion below).

Quantitative information on the simulation accuracy of the improved and the modified PELMO versions referring to the wind-tunnel findings is summarized in Table 2. A simple linear regression between the sets of values yields the following. For chlorpyrifos, fenpropimorph, and parathion-methyl, the r2 values of the improved PELMO version as a measure of how successful the regression was in explaining the results were closer to 1, thus representing a closer correlation. With the exception of terbuthylazine, the best-fit lines of the improved model calculations were more precise than the calculations using the previous PELMO version. For all compounds, negative slopes were calculated when using the previous model version. Obviously, previous model predictions change in the direction opposite to the measurements. The slopes of the linear regression of the improved calculations on measured values being positive in all cases indicate changes in the same direction, thus suggesting a more accurate simulation. The Y intercepts of the regression lines of the improved calculations were closer to 0 than those corresponding to previous calculations, thus also suggesting that the application of the improved model results in a better match with wind-tunnel findings. However, as a simulation that fits measured values perfectly would have a slope, intercept, and r2 of 1, 0, and 1, respectively, variations indicate the need for further calibration of the model approach.


View this table:
[in this window]
[in a new window]
  		Table 2. Linear regression for comparing the simulation accuracy of the official and the improved Pesticide Leaching Model (PELMO) version based on wind-tunnel findings in the range between Days 1 and 13.

 
Influence of Henry's Law Constants on Model Predictions
In addition to the requirement of an improved physical-based description, the reliability of model predictions generally depends on the quality of the underlying data and is largely determined by the extent to which model input parameters can be inferred from readily available data (Walden and Haith, 2003). Due to the volatilization rates being calculated on the basis of Henry's law (Eq. [1]), the soil–water partitioning coefficient (Henry's law constant H') is the most important pesticide property affecting PELMO's predictions. The significant effect of H', including the variation of default values within the range of literature values on calculated volatilization rates of parathion-methyl (H' between 5.2 x 10–7 and 3.4 x 10–6) and chlorpyrifos (H' between 1.6 x 10–4 and 2.7 x 10–4), is exemplified in Fig. 3. For both compounds, a marked increase in volatilization was calculated and a disproportionate effect of irrigation on volatilization rates was predicted after H' was enhanced (Fig. 3). With regard to chlorpyrifos, an apparently slight increase of H' from 1.6 x 10–4 to 2.7 x 10–4 resulted in a marked increase of predicted volatilization (Fig. 3B), especially under moist conditions at the early stage of the simulation and after irrigation was given (Day 8). Taking into consideration that surface-applied compounds generally show the highest volatilization rates during the first hours after application, this increase obviously affects significantly the predicted cumulative volatilization. Thus, uncertainty in experimental determination of H' is a crucial factor influencing predicted volatilization rates, indicating that reliable experimental procedures for measurement of key factors (e.g., H') are as important as accurate model approaches for reliable computations. Various experimental methods are available for measurement of Henry's law constants (e.g., Rice et al., 1997; Fendinger and Glotfelty, 1988; Mackay et al., 1979) based on the determination of the respective pesticide concentrations in water and air in an equilibrated system. All of these methods need great care to achieve reproducible results and their applicability to low-volatile compounds is restricted due to analytical detection limits. Particularly with regard to the quantification of temperature dependence of water–air partitioning, which would be of importance for modeling pesticide volatilization, the analytical limitations make reliable measurements difficult.



View larger version (26K):
[in this window]
[in a new window]
  		Fig. 3. Calculations using the improved PELMO version, including variation of Henry's law constants (dimensionless values, calculated at 25°C). (A) Parathion-methyl. (B) Chlorpyrifos.

 
Field Study: Experimental Findings and PELMO Calculations
A comparison between predictions of previous and improved PELMO versions and measured volatilization rates of parathion-methyl, fenpropimorph, and terbuthylazine obtained in a 20-d field study is given in Fig. 4. For parathion-methyl and fenpropimorph, volatilization rates up to 90 µg m–2 h–1 (Fig. 4A, 4B) were detected immediately after application; during the following days volatilization rates exhibited a sustained decline (volatilization rates of fenpropimorph fall below the detection limit) and increased markedly after irrigation on Day 6 and to a lower degree after irrigation on Day 15 (Fig. 5). Compared with fenpropimorph and parathion-methyl, volatilization rates of terbuthylazine remained on a lower level during the first days but increased sharply after irrigation (Fig. 4C). Although the kinetics of both field and wind-tunnel results followed approximately the same pattern, the effect of rainfall on volatilization determined in the field study exceeded the irrigation effect measured in the wind-tunnel experiment. This is clearly attributed to the different soil moisture profiles over the course of the studies. During the first week after application, a drying out of the lysimeter used in the wind-tunnel study was observed, illustrated by soil moisture falling below 10%vol (Fig. 1A). Obviously, irrigation of 8 mm on Day 8 did not sufficiently compensate for the water loss of the top soil layer and caused only a slight increase of soil moisture and volatilization rates during the coming day. Measurements within the field study (Fig. 5) revealed soil moisture to markedly exceed the corresponding wind-tunnel findings and rainfall to cause a much higher increase of the water content of the top soil. Consequently, a significant increase of the volatilization rates of the applied compounds was detected in the field after intense rainfall events on Day 6.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 4. Measured (field study) and predicted (PELMO) volatilization fluxes after soil surface application to orthic luvisol. (A) Parathion-methyl. (B) Fenpropimorph. (C) Terbuthylazine.

 


View larger version (35K):
[in this window]
[in a new window]
  		Fig. 5. Measured and calculated soil moisture profiles over the course of the field study on orthic luvisol using the official PELMO version (FOCUS PELMO, top compartment thickness: 5 cm) and the modified version (top compartment thickness: 1 mm).

 
The PELMO calculations were performed on an hourly basis to enable a better temporal resolution of increased volatilization rates immediately after application. For the consideration of a period of equilibration, diminished sorption coefficients during the first three hours after application were used in the improved PELMO version. For fenpropimorph and parathion-methyl, PELMO predictions at the initial stage of the simulation agreed reasonably well with experimental findings, whereas volatilization rates of terbuthylazine were overestimated. In accordance with measurements, the model predicted enhanced volatilization rates after irrigation was given. Indeed, this effect was underestimated for the low-sorbing compounds parathion-methyl and terbuthylazine, but slightly overestimated for fenpropimorph. The different extent of the irrigation effect is due to low-sorbing compounds parathion-methyl and terbuthylazine being translocated below the 1-mm top soil compartment after rainfall events. As PELMO assumes volatilization to occur exclusively from the top layer, the translocation after irrigation causes a lower increase of volatilization of terbuthylazine and parathion-methyl in comparison with the strongly sorbed fenpropimorph (see Table 1), which remained mainly in the top layer. Calculations revealed that about 64% of the soil deposit of terbuthylazine was transported below the top millimeter of the soil after rainfall events on Day 6, whereas only approximately 8% of fenpropimorph was displaced and thus kept from being volatilized.

For fenpropimorph, PELMO predicts the effect of rainfall on Day 15 to exceed the corresponding effect after rainfall on Day 6 resulting in calculation of higher volatilization fluxes on Day 16 than on Day 7, which is inconsistent with experimental findings (Fig. 4B). Obviously, the computation of increased volatilization fluxes on Day 16 is mainly attributed to the above-mentioned high sorption, which leads to nearly constant fenpropimorph concentrations in the soil top layer over the course of the study. In addition, increasing air temperature at the end of the study (approximately 20°C, data not shown) led to a significant increase of effective vapor pressure of the applied compounds. The main reason for the contradiction between PELMO calculation and the experimentally determined volatilization flux after Day 15 appears to be the difference between predicted and measured soil moisture. Even though the new PELMO version enabled a better reflection of the measured soil moisture at a depth of 50 mm than the official PELMO version (Fig. 5), the effect of irrigation on calculated water content of the top millimeter of the soil is still uncertain. For rainfall events on Days 6 and 15, PELMO predicts a marked increase of soil moisture, respectively. Due to the correlation between water content and volatilization fluxes included in the improved PELMO version, enhanced volatilization fluxes were calculated. However, measurements revealed that the increase of soil moisture at the end of the study calculated by PELMO, especially after irrigation on Days 15 and 18, was overestimated and hence volatilization fluxes were also overestimated. The situation is made even more complex by the calculation of the water content of the top millimeter of the soil, which is beyond experimenter's control. An increase of the water content of the top layer up to approximately 35%vol on Day 15 and a slow decrease during the following 2 d, as predicted by PELMO, is doubtful, especially with regard to the increased air temperature at the end of the study, which facilitates a fast drying of the top soil. Due to the significant effect of the water content of the top layer on volatilization from the soil surface, this uncertainty is a general limitation of volatilization predictions.

Thus, a major task for future model improvement will be a reliable estimation of the water content in the top millimeters of the soil and the subsequent calibration of new model approaches. This calibration is dependent on the availability of high-resolution data sets. Currently available data sets based on experimental volatilization studies performed at the field or semi-field scale do not provide sufficient information on water content and concentrations of pesticide and formulation additives in the top millimeter of the soil. For that reason, advanced experimental procedures for a deeper understanding of the underlying processes governing the extent of volatilization are required. For instance, it is well known that near-surface water content can be measured by visible and thermal infrared remote sensing, as well as active and passive microwave remote sensing techniques. Whether or not these techniques will be applicable within future studies on volatilization for measuring the spatial distribution and temporal variation of soil moisture content will be dependent on their resolution and the noise sources within the experimental scenarios.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The implementation of soil-moisture-dependent soil–water partitioning coefficients, the reduction of the compartment size of the top soil layer, and the consideration of temperature-dependent Henry's law constants in an improved PELMO version resulted in a much better agreement between computations and measurements. For both semi-field (wind tunnel) and field studies, the improved PELMO version reflected the time course of volatilization much better than the official PELMO version, especially at the initial stages of the studies. Due to the extent of volatilization being significantly influenced by the thickness of the top soil layer, further studies on pesticide and water distribution in the top layer are still required.

The prediction of enhanced volatilization fluxes for increasing water content was in accordance with experimental findings and represented a considerable improvement of the official PELMO version. However, the extent of the predicted volatilization increase after irrigation was still afflicted with uncertainty and illustrated the need for further calibration. The increase of sorption coefficients by a correction factor of up to 100 below the wilting point is to be replaced by experimentally determined sorption coefficients at low water content. Furthermore, the broad range of Henry's law constants given in literature was shown to have a major effect on model calculations, thus also illustrating the need for enhanced techniques to measure key parameters in support of model development.

The presented approach will enable a more reliable computation of volatilization from bare soil in future PELMO calculations. Main limitations of the approach are clearly attributed to a lack of knowledge on main processes affecting the fate of pesticides at the soil–air interface. Based on currently available experimental findings on volatilization, such a deeper understanding is hard to obtain. Therefore, advanced modeling to improve the presented boundary-layer concept will be very closely connected with advanced experimentation and the development of suitable techniques for elucidating phase distribution at the top layer of the soil.


   ACKNOWLEDGMENTS
 
Funding was provided by the European Commission within the framework of the APECOP project (Effective Approaches for Assessing the Predicted Environmental Concentrations of Pesticides). The authors wish to thank BASF AG (Ludwigshafen, Germany) and Syngenta AG (Basel, Switzerland) for providing 14C-labeled compounds.


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


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2004 33: 1589-1599. [Full Text]  




This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
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 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 Google Scholar
Google Scholar
Right arrow Articles by Wolters, A.
Right arrow Articles by Vereecken, H.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wolters, A.
Right arrow Articles by Vereecken, H.
Agricola
Right arrow Articles by Wolters, A.
Right arrow Articles by Vereecken, H.
Related Collections
Right arrow Pesticides
Right arrow Organic Compounds
Right arrow Other Models
Right arrow Agricultural Pesticides
Right arrow Air 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