Environ. Sci. Technol. 2004, 38, 2885-2893
Pesticide Volatilization from Plants: Improvement of the PEC Model PELMO Based on a Boundary-Layer Concept A N D R EÄ W O L T E R S , * , † M I N Z E L E I S T R A , ‡ VOLKER LINNEMANN,† MICHAEL KLEIN,§ A N D R E A S S C H A¨ F F E R , § A N D HARRY VEREECKEN† Forschungszentrum Ju ¨ lich GmbH, Institute of Chemistry and Dynamics of the Geosphere IV: Agrosphere, 52425 Ju ¨ lich, Germany, Alterra Green World Research, Wageningen University and Research Centre, P.O. Box 47, 6700 AA Wageningen, The Netherlands, and Fraunhofer-Institute for Molecular Biology and Applied Ecology, P.O. Box 1260, 57377 Schmallenberg, Germany
Calculation of pesticide volatilization from plants as an integral component of pesticide fate models is of utmost importance, especially as part of PEC (predicted environmental concentrations) models used in the registration procedures for pesticides. A mechanistic approach using a laminar airboundary layer concept to predict volatilization from plant surfaces was compared to data obtained in a windtunnel study after simultaneous application of parathionmethyl, fenpropimorph, and quinoxyfen to winter wheat. Parathion-methyl was shown to have the highest volatilization during the wind-tunnel study of 10 days (29.2%). Volatilization of quinoxyfen was about 15.0%, revealing a higher volatilization tendency than fenpropimorph (6.0%), which is attributed to enhanced penetration of fenpropimorph counteracting volatilization. Predictions of the boundarylayer approach were markedly influenced by the selected values for the equivalent thickness of the boundary layer and rate coefficients, thus indicating that future improvements of the approach will require a deeper understanding of the kinetics of the underlying processes, e.g. phototransformation and penetration. The boundarylayer volatilization module was included in the European registration model PELMO, enabling simultaneous calculation of volatilization from plants and soil. Application of PELMO to experimental findings were the first comprehensive PEC model calculations to imply the relevant processes affecting the postapplication fate of pesticides.
Introduction Volatilization of pesticides from soil and plant surfaces and subsequent atmospheric transport is a primary pathway by which pesticides may be dispersed throughout the general environment and may lead to contamination by long-range transport and deposition at locations remote from their application (1, 2). * Corresponding author phone: +49-2461-618668; fax: +49-2461612518; e-mail:
[email protected]. † Forschungszentrum Ju ¨ lich GmbH. ‡ Alterra Green World Research. § Fraunhofer-Institute for Molecular Biology and Applied Ecology. 10.1021/es035061m CCC: $27.50 Published on Web 04/15/2004
2004 American Chemical Society
Main factors affecting volatilization of pesticides from crops are physicochemical properties of the compounds, the persistence on the plant surface and environmental conditions. Persistence on the leaf surface depends on various dissipation processes, such as photodegradation, wash-off from the leaves, and penetration into the plant leaves (3). Recent studies on plant-air transfer of xenobiotics mainly focus on the pathways into the leaves and possible storage locations to elucidate the implications for mechanisms and modeling (4). One of the most important differences between volatilization from soil and plants is due to the varying aerodynamic scenarios. Simulations showed that under convective conditions the vapor emitted by a source point rapidly forms stationary envelopes around the leaves. Vapor concentrations within these unstirred layers depend on the vapor pressure of the compound in question and on its affinity to the lipoid surface layers on the leaf (5). Despite intense research in recent years, including the development of numerous laboratory and field methods to measure volatilization rates (6-9), knowledge of ratedetermining processes is currently not sufficient for developing a reliable, physically based model approach to predict volatilization fluxes from plant surfaces. As a screening-level approach for estimating the initial volatilization rates after application to crops, a correlation between the logarithm of the volatilization rates and the logarithm of pesticide vapor pressures is used (10). However, comprehensive approaches, as already developed for the prediction of volatilization from soil (11-14), are still not available. Precise simulation of volatilization as an integral component of a complete pesticide transport model is of highest importance. Attention should be paid to the fact that many models used for pesticide registration, including those used by USEPA, do not take volatilization losses into account when evaluating surface water and groundwater contamination risks (15, 16). Whereas the European registration models PELMO (pesticide leaching model) and PEARL (pesticide emission assessment at regional and local scales (17)) already include simplified approaches for the prediction of volatilization from soil (18, 19), volatilization from plants is still not sufficiently considered. To describe volatilization in a more mechanistic way, a simplified model on the basis of a series of well-defined experiments using the fungicide fenpropimorph was developed (20, 21). This model takes into account the most important factors influencing volatilization from plants. This paper summarizes results on the volatilization of three pesticides (14C-labeled parathion-methyl, nonlabeled fenpropimorph, and quinoxyfen) determined in a windtunnel study of 10 d after simultaneous application to winter wheat. Measurements were compared with the output of the novel boundary-layer approach. Implementation of the approach in the PEC model PELMO enabled a comprehensive survey of the main processes governing the environmental distribution of the pesticides after application, including volatilization from soil and plants.
Experimental Section Wind Tunnel. A glass wind tunnel was set up above a lysimeter with a soil surface area of 0.5 m2 to measure the gaseous emissions of the applied pesticide mixture (Figure 1). A single blower presses air into the wind tunnel after intensive cleaning by various filter stages (max. 1500 m3 h-1). Air filters and subsequent sieves guarantee a uniform air stream through the glass tunnel. The top of the wind tunnel can be adjusted in height and thus be adapted to the level of growing plants. In this way, constant wind velocities from VOL. 38, NO. 10, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2885
FIGURE 1. Schematic of the wind tunnel for measuring volatilization of pesticides from winter wheat under fieldlike conditions. AF ) active charcoal filter, B ) brine tank, C ) cooler, CV ) converter, DA ) data acquisition, FF ) fine filter, GF ) glass fiber filter, HVS ) high-volume sampler (14C-labeled organics), MVS ) medium-volume sampler (14C-labeled carbon dioxide), P ) pump/blower, PF ) prefilter, PUF ) polyurethane foam, R ) refrigeration, TDR ) time domain reflectometry, XAD ) adsorbing resins (Amberlite XAD-4). 0.3 to 3.5 m s-1 can be achieved with a minimum air volume. Fieldlike conditions are simulated inside this wind tunnel by a continuous, automatic adjustment of the air temperature to the outdoor situation. The use of UV-transparent glass (side walls) and UV-transparent acrylic glass (lid) as construction materials guarantees for sufficient irradiation and light quality. During the experiments, basic climatic parameters were monitored using various sensors, including the determination of volumetric water content in the soil by time domain reflectometry (TDR) measurement at various depths (Figure 2A). Precipitation events were simulated by irrigation nozzles in the lid of the wind tunnel. A detailed description of the system was given by Stork et al. (8). Application Device, Compounds, and Experimental Soil. Three pesticides, covering a broad range of physicochemical properties (Table 1), were applied using a semiautomatic sprayer unit. The pressure tank of the spray mixture was flow-optimized to guarantee bubble-free and homogeneous spraying of the fluid. For application, the sprayer was placed above the wind tunnel. A contamination shield lined with aluminum foil was provided inside. All application losses were detected by decontamination to determine the net applied radioactivity. Application conditions, including application rate and volume of applied water, were chosen in accordance with agricultural practice (Table 1). Experimental soil was classified as an orthic luvisol, and the Ap horizon (plow layer) contained 1.1% Corg, 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 the wilting point was 14.9%vol. Sampling. Organic compounds in the exhaust air were sampled with a high-volume sampler consisting of an adhesive-free glass fiber filter (185-mm o.d.) to trap particulate matter followed by three polyurethane foam plugs (100-mm o.d. × 150-mm length) held within a glass sleeve. Aliquots were taken isokinetically based on industrial guidelines for the sampling of stack air. The maximum sampling rate was 50 m3 h-1, corresponding to 3-10% of the total airflow through the wind tunnel. Sampling period of the highvolume sampler was between 1 h to a maximum of 24 h. A 2886
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 10, 2004
FIGURE 2. Climatic conditions over the course of the wind-tunnel experiment. A: Volumetric water content measured by TDR equipment at several depths. B: Radiation and soil surface temperature. medium-volume sampler was used at a sampling rate of 1.03.5 L min-1 over a maximum sampling interval of 48 h each for measuring 14CO2 arising from the mineralization of 14Clabeled fenpropimorph. The sample gas was dried by silica
TABLE 1. Physicochemical Active Ingredient Data of the Investigated Compounds (28) and Application Details
a
Radiochemical purity: > 99%.
b 14C-labeled:
33.4% (15.56 MBq).
gel and phosphorus pentoxide. 14CO2 was subsequently absorbed by 2-methoxy-propylamine using a special cooled intensive wash bottle (22). Losses of highly volatile adsorbent were minimized by intensive reflux cooling (-40 °C). At the end of the wind-tunnel study, plants were completely harvested and rinsed with solvents of decreasing polarity (2-3 L of water, methanol and chloroform, respectively). Soil layers up to 10 cm were completely removed, and soil samples of deeper layers were taken. Leachate was also sampled at the end of the study. Extraction and Analytics. Foam plugs were extracted separately with methanol using a special squeezing apparatus (23). Extracts were reduced to 5-10 mL using a rotary evaporator and further concentrated by nitrogen blow-down. Glass fiber filters and soil samples were extracted with methanol in a Soxhlet apparatus for 16 h. Nonextractable radioactivity in soil material was measured by combustion. Radioactivity of extracts was determined by liquid scintillation counting. Active ingredients of air, plants, and soil samples were characterized by radio-high performance liquid chromatography and radio-thin-layer chromatography. Nonlabeled compounds were quantified with a Hewlett-Packard 6890 gas chromatograph (GC) equipped with a split inlet, 30-m fused silica DB-1 capillary column with a 0.25-mm i.d. and a 0.25µm film thickness, and a mass-selective detector (MSD). The MSD source (held at 230 °C) was operated in electron ionization mode, while the mass filter quadrupole (held at 150 °C) was operated in selective ion monitoring mode. Selected ions for fenpropimorph (m/z: 128, 173, 303) and quinoxyfen (m/z: 197, 199, 314) allowed for the quantification of target compounds. The injector and GC-MSD transfer line were operated at 240 and 270 °C, respectively. An HP 6890 series autoinjector was used to inject 1 µL of sample in splitless mode. The oven temperature started at 90 °C (3-min hold) and was programmed at 8 °C/min to 270 °C and held for 1 min. Throughout the run the carrier gas (helium) was maintained at 1.0 mL min-1. Pesticide concentrations were calculated by comparing the ratio of peak area response of
the analyte over the internal standard (1-methylpyrene) in samples with those of calibration standards.
Theoretical Considerations Volatilization from Plants (Boundary-Layer Concept). A basic approach for predicting the environmental fate of pesticides after application to plant surfaces was used, including volatilization from the leaves, penetration into the leaves, wash-off, and phototransformation. Well-known processes, e.g. the calculation of volatilization fluxes by determining vapor diffusion through the laminar air boundary layer, were described mechanistically. Other processes and factors, especially photochemical transformation and penetration into the leaves, were described using an empirical approach. This empirical estimation requires the equivalent thickness of the air boundary layer and the rate coefficients for penetration, phototransformation, and wash-off to be calibrated. The net applied amount of pesticide and the fraction of dosage intercepted by the crops were introduced in the model. The vapor pressure of the pesticide at the prevailing temperature is calculated by an adapted form of the Clausius-Clapeyron equation (24). The concentration of the pesticide in the air at the deposit surface on the leaves is calculated by using the general gas law
Ca,s )
( ) MPact RT
(1)
where Ca,s ) pesticide concentration in the air at the plant surface [kg m-3], M ) molar mass [kg mole-1], and Pact ) actual vapor pressure of the pesticide [Pa]. The coefficient Da [m2 d-1] for pesticide diffusion in air at the prevailing temperature is calculated according to Tucker and Nelken (25). Volatilization of pesticide from the deposit surface on the leaves is determined by vapor diffusion through the laminar air boundary layer. The potential rate of volatilization of pesticide from the deposit/leaf surface is calculated VOL. 38, NO. 10, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2887
dAp ) -Jvol,act - Rpen - Rw - Rph dt
as follows
(Ca,s - Ca,t) Jvol,pot ) Da dlam
(2)
where Jvol,pot ) potential flux of volatilization from the surface [kg m-2 d-1], Ca,t ) concentration in the turbulent air just outside the laminar air layer [kg m-3], and dlam ) equivalent thickness of the laminar air boundary layer [m]. The pesticide concentration in the turbulent air layer outside the laminar boundary layer (Ca,t) is taken to be zero. The actual rate of pesticide volatilization is described by taking into account the mass of pesticide deposited on the plants, the distribution in the canopy and the formulation of the pesticide
Jvol,act ) fmas fdis ffor Jvol,pot
(3)
where Jvol,act ) actual rate of pesticide volatilization [kg m-2 d-1], fmas ) factor accounting for the effect of pesticide mass on the plants [-], fdis ) factor accounting for the effect of pesticide distribution in the canopy [-], and ffor ) factor accounting or the effect of formulation [-]. The effect of spray distribution in the crop canopy on volatilization (fdis) is considered to be 1.0 (standard distribution). The factor considering the effect of the formulation on volatilization (ffor) is assumed to be 1.0 for spraying standard formulations, e.g. emulsifiable concentrates and wettable powders. The pesticide is assumed to be deposited on the leaves in spots with variable thickness. The thinner the deposit at a certain place is, the earlier that spot will be depleted by volatilization. The concept is based on the assumption that the volatilizing surface decreases in proportion to the decrease in mass of pesticide in the deposit. So, the factor fmas is calculated from the ratio Ap/Ap,ref, where Ap ) areic mass of pesticide on plants [kg m-2], and Ap,ref ) reference areic mass of pesticide on plants [10-4 kg m-2] () 1 kg ha-1). The rate of pesticide penetration into the leaves is calculated by
Rpen ) kpenAp
(4)
where Rpen ) rate of pesticide penetration into the leaves [kg m-2 d-1], and kpen ) rate coefficient of penetration [d-1]. The rate of pesticide wash-off from the leaves by rainfall is set dependent on rainfall intensity and wash-off coefficient
Rw ) kw Wr Ap
(5)
where Rw ) rate of pesticide wash-off from the leaves [kg m-2 d-1], kw ) coefficient for pesticide wash-off [mm-1], and Wr ) rainfall intensity [mm d-1]. The rate of pesticide transformation by solar irradiation is described by first-order kinetics
Rph ) kphAp
(6)
where Rph ) rate of phototransformation on the leaves [kg m-2 d-1], and kph ) rate coefficient of phototransformation [d-1]. The rate coefficient kph is set dependent on sunlight irradiation intensity
kph )
( )
Iact k Iref ph,ref
(7)
where Iact ) actual solar irradiation intensity [W m-2], Iref ) reference solar irradiation intensity [500 W m-2], and kph,ref ) rate coefficient of phototransformation at reference irradiation intensity [d-1]. The equation for the conservation of mass of pesticide on the plants reads as follows: 2888
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 10, 2004
(8)
The model provides the option to distinguish two deposit classes: a well-exposed and a poorly exposed class. The deposit in the latter class may be enclosed by plant parts (e.g. in leaf axils), it might be located on the lee side of the airflow, or it is assumed to be located deeper in the canopy. To trace the effect of two deposit classes on the course of volatilization, the fraction of poorly exposed deposit was set at 20% of the applied dose. Furthermore, the rates of the decrease processes for the poorly exposed deposit were set at 20% of the rates of the corresponding processes for the well-exposed deposit. The definition of two deposit classes requires the use of two mass conservation equations, one for the well-exposed and one for the poorly exposed deposit. A detailed description of the model approach is given by Leistra and Wolters (20). Implementation in PEC Model PELMO. The boundarylayer concept was included in PELMO, enabling the simultaneous calculation of volatilization from plants and soil. For the calculation of the soil processes, including volatilization, degradation, and root-uptake, an advanced module was applied to the soil deposit (18). The following equation is used for the calculation of the volatilization rates
H′csol d
J ) -D
(9)
where J ) volatilization rate [g cm-2 d-1], D ) diffusion coefficient in air [cm2 d-1], H′ ) nondimensional Henry’s law constant [-], d ) air boundary layer [cm], and csol ) 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 air boundary layer (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 (26) was improved, and the following modifications of the volatilization module were implemented: Based on Henry’s law constants measured or estimated at two temperatures, exponential approaches were included for calculating the temperature-dependence of water-air partitioning over the relevant temperature range. This enables PELMO to assess Henry’s law constants over the course of the study for actual soil temperatures. A small compartment of 1 mm was added on top of the soil for a more realistic reflection of the pesticide concentration at the soil surface. As an additional effect, the lower top compartment size results in changing soil moisture of the top millimeter over the course of the study. The moisture dependence of the soil adsorption coefficients at low water content was taken into consideration. For moistures ranging from above air-dry conditions to below wilting point, an increase of the soil sorption coefficient was assumed to occur in the top millimeter of the soil. The sorption coefficient at 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. The sorption coefficient at moisture contents between air-dry and that corresponding to wilting point is obtained by interpolation using a power function
Kd,eff ) Kd‚Fss for Θ < Θad
(10)
Kd,eff ) Kd‚10m for Θad < Θ < Θwp
(11)
where Fss ) increase of soil sorption when soil is air-dry [-], Θ ) volume fraction liquid phase [m3 m-3], Θwp ) volume fraction liquid phase at wilting point [m3 m-3], and Θad ) volume fraction liquid phase when soil is air-dry [m3 m-3]. Above the wilting point, the sorption coefficient remained
TABLE 2. Radioactivity Balance of 14C-Labeled Parathion-Methyle sample
radioactivity [% AR]
contaminationa
NDb
soil extractable nonextractable leachate control plot plants leaves roots volatilization parent compound metabolites mineralization to 14CO2 glass fiber filter (particulate matter) Σ14C
4.4c 13.6d NDb 0.1 21.5 0.5 29.2 3.3 6.4 8.2 87.2
a Contamination of the wind tunnel and the high-volume sampler. ND ) not detectable. c Soil layer (0-2.5 cm): 4.1%AR; soil layer (2.55.0 cm): 0.3%AR. d Soil layer (0-2.5 cm): 13.0%AR; soil layer (2.5-5.0 cm): 0.6%AR. e Net applied radioactivity (AR) ) 100%. b
unchanged. The exponent m is given by
m)
[
]
(10log Θ - 10log Θwp) log Fss 10 ( log Θad - 10log Θwp)
10
(12)
Results and Discussion Radioactivity Balance and 14C-Labeled Residues. Recovery of 14C-labeled parathion-methyl was about 87.2% AR (Table 2), thus approximating the guideline values for registration purposes ranging from 90 to 110% of the applied amount. Both, radioactivity of the leachate and system contaminations were below detection limits. After finishing the experiment, the major part of the remaining radioactivity was detected on the plants, including extractable and nonextractable fractions of leaves and roots (approximately 22% AR). In comparison with a previous wind-tunnel study on bare soil (approximately 76% AR characterized as soil residues (19)) only small amounts of soil residues were detected (18% AR). These findings are attributed to the lysimeter being abundantly covered with vegetation during application. In agreement with the wind-tunnel study on bare soil, a displacement of radioactivity in deeper layers (> 5 cm) did not occur. About 13.6% AR remained in the soil and was not Soxhletextractable, indicating increasing sorption of parathionmethyl on the soil during the experiment. In previous studies, the glass fiber filter was shown to have a retention of > 99.9% for particles > 10 nm on a filter test apparatus (8). It may thus be assumed that all particulate residues in air are quantitatively retained (8.4% AR). However, it is to be considered that adsorption of gaseous residues on the glass fiber filters over the course of the experiment cannot be excluded since many pesticide vapors possess high adsorptive properties (27). Therefore, a clear distinction
between gaseous and particulate residues adsorbed on the filters is not possible. Besides unchanged parathion-methyl, 17.4% of the radioactivity detected in methanol extracts of the PUF plugs was characterized as metabolites, suggesting that metabolic transformation on the plant surfaces might have contributed to the formation of volatile metabolites which were subsequently trapped on the PUF plugs (Table 3). A high quantity of metabolites (83% of extracted radioactivity) was detected in the methanol extracts of the glass fiber filters used for the retention of particulate residues in air. The magnitude of metabolization exceeds the corresponding amount of metabolites detected in the soil samples (71.5% of extracted radioactivity). Metabolization of parathion-methyl on the PUF plugs or glass fiber filters during sampling, as suggested by Stork et al. (8), is considered as an additional degradative pathway which might have caused the higher amounts of metabolites in air samples compared to soil samples. Nonlabeled Residues. Due to the use of nonlabeled compounds, no complete mass balances were obtained for fenpropimorph and quinoxyfen, especially the amount of mineralization to 14CO2 was not quantitated. Distribution of both fenpropimorph and quinoxyfen in the different solvents after rinsing the plants showed the highest amounts of the pesticides in the methanol and chloroform samples, whereas only slight amounts of quinoxyfen and no fenpropimorph were identified in the aqueous solutions (Table 4). Due to its mode of action as systemic pesticide, fenpropimorph obviously penetrates into the leaves enabling a long-term protective effect. Quinoxyfen is known to be extensively photodegraded on the wheat leaf surface, whereas it is only slightly metabolized in wheat, with low residues found in crops (28). Therefore, the slight amount of quinoxyfen in the aqueous solution and the comparably high recovery of quinoxyfen in the methanol and chloroform phases may be attributed to the photodegradation occurring at the plant surface. Metabolites of fenpropimorph and quinoxyfen were not targeted in this study. Soil and plant residues revealed higher amounts of quinoxyfen in comparison with fenpropimorph, indicating a lower metabolization of quinoxyfen over the course of the experiment. These findings were in agreement with the minimal photolysis of quinoxyfen on soil surfaces (DT50 > 1 year, Southern England) ascertained in previous studies (28). Fenpropimorph was shown to form large amounts of nonextractable residues and metabolites, e.g. fenpropimorph acid (29). So, the small amount of detectable residues is probably due to considerable metabolization of fenpropimorph. Volatilization Measurements. In total, 32.5% AR were volatilized by the end of the wind-tunnel study, corresponding to 29.2% parathion-methyl and 3.3% of metabolites (Table 2). About 25% AR already volatilized within the first 24 h, followed by a sharp decrease (Figure 3A), indicating other processes such as foliar uptake counteracting the volatilization process. The major part of volatilization took place immediately after application due to a weak adsorption/ bonding of the pesticide on the leaf surface. Decreasing
TABLE 3. 14C-Labeled Compounds in the Methanol Extracts of Soil Samples, PUF Plugs, and Glass Fiber Filtersc
a
compound
soil layer (0-2.5 cm)
PUF plugsa
glass fiber filter
parathion-methyl 4-nitrophenol paraoxon-methyl parathion-methyl-S-nitrophenylester unknown polar products
28.5 34.1 NDb NDb 37.4
82.6 4.1 4.1 0.8 8.4
17.0 42.5 6.8 NDb 33.7
PUF: polyurethane foam (values averaged within the experimental duration). b ND ) not detectable. c Data in % of extracted radioactivity.
VOL. 38, NO. 10, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2889
FIGURE 3. Volatilization after application to winter wheat. A: Cumulative volatilized amount. Net applied active ingredients ) 100%. B: Volatilization fluxes.
TABLE 4. Soil and Plant Residues of Quinoxyfen and Fenpropimorph after Finishing a Wind-Tunnel Study on Winter Wheatb
a
sample
fenpropimorph
quinoxyfen
soil residues plant residues water methanol chloroform
2.0
5.4
NDa 0.7 0.9
0.9 4.9 5.1
ND ) not detectable.
b
Data in % of net applied amount.
volatilization rates were observed following the first 24 h, illustrating the long-term release of protected or adsorbed pesticide residues. Increasing air temperatures and irrigation caused only temporary increases of volatilization rates. Due to a high soil temperature (approximately 27.5 °C; Figure 2B), the volatilization rate of parathion-methyl increased at the beginning of day 4 and decreased subsequently parallel to a decrease of temperature. A sharp rise of volatilization was observed on day 7 and day 8, which may be attributed to simultaneous irrigation. The effect of irrigation was enhanced by an additional increase in temperature on day 7. In total, 6.0% of the applied fenpropimorph volatilized in the wind tunnel (Figure 3A). Only about 9.6% of the applied amount of fenpropimorph, including soil and plant residues and the volatilized fraction, was recovered after 10 days. One possible explanation is a rapid decomposition of fenpropimorph under semifield conditions due to the UV transparency of the side walls of the tunnel facilitating photochemical degradation. Indirect methods (measurement of residues) showed 70% of the applied fenpropimorph or its metabolites 2890
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 10, 2004
had been lost undetected via the atmosphere, as determined in a lysimeter study on winter wheat (30), while measurements of volatilization including the detection of metabolites revealed cumulative volatilization up to 60% of the applied radioactivity from summer barley (31). Taking into consideration that metabolites were not analyzed within the study on winter wheat, the results imply that a major portion of the applied fenpropimorph might have volatilized as metabolites. As an additional process counteracting volatilization, penetration into the leaves has to be considered. The results of soil and plant analysis (Table 4) revealed that major parts of the applied amount of fenpropimorph had penetrated into the crops over the course of the study, indicating that both penetration and metabolization had contributed to the low volatilization of the parent compound. Besides negligible deviations on day 1, the volatilization kinetics of fenpropimorph followed roughly the same course as parathion-methyl (Figure 3B). Volatilization rates showed a marked decrease on day 9, which may be attributed to the disappearance of fenpropimorph from the plant surfaces, illustrated by a nondetectable fenpropimorph amount in the aqueous wash solution (Table 4). Cumulative volatilization of quinoxyfen over the course of the experiment was about 15.0% of the net applied amount (Figure 3A), apparently indicating a higher volatilization tendency of quinoxyfen in comparison with fenpropimorph. However, due to the lower metabolization of quinoxyfen, which is documented by a recovery rate of approximately 31.0%, the lower cumulative volatilization of fenpropimorph may be ascribed to its higher degradability, the occurrence of nondetectable metabolites, and its systemic activity. The low application dose of quinoxyfen (65.38 g ha-1; Table 1) in comparison with fenpropimorph (247.87 g ha-1) accounts for the volatilization rates of quinoxyfen falling below the corresponding values of fenpropimorph during the wind-tunnel study (Figure 3B). However, a clear dependence of volatilization of quinoxyfen on air temperature and irrigation, illustrated by increasing rates on day 4 and day 7, was established. Calculations with Boundary-Layer Concept. Adjusting the boundary-layer thickness and the rate coefficients for penetration and phototransformation to the experimental findings, as shown in Figure 4A under well-exposed conditions, revealed a reasonable agreement between experimental findings and predictions for parathion-methyl. However, the initial volatilization rates during the first hours of the experiment were significantly underestimated by the model approach. The predicted disappearance of detectable residues on the leaves after 3.6 days and the subsequent disrupting of volatilization is contrary to experimental findings, which revealed slight volatilization rates by the end of the study. In addition, measurements of the rinsability of the pesticides after finishing the experiment by using solvents of decreasing polarity revealed that about 1.2% of the net applied parathion-methyl were rinsable with water, indicating that the prediction of a complete disappearance of parathionmethyl during the first 2 days of the study required revision. Replacing the well-exposed scenario by a poorly exposed scenario led to a more realistic reflection of the time course of the disappearance of the plant deposit at the later stage of the study (Figure 4B). Even though this modification generally improved the agreement of observed and simulated volatilization of parathion-methyl at the end of the study, the underestimation of volatilization during the first day increased. Reducing the deposit being available for volatilization by introducing a class of plant deposits being poorly exposed causes a decrease in volatilization rates at the beginning. Obviously, the lower volatilization at the beginning is compensated by higher rates at the end, finally leading to comparable cumulative volatilization under well-exposed and poorly exposed conditions over the total experimental period.
FIGURE 4. Measured and predicted cumulative volatilization of parathion-methyl. A: Well-exposed scenario. B: Poorly exposed scenario. For both scenarios, a diffusion coefficient in air of 0.5 m2 d-1 was used, and the following values for the parameters to be calibrated were obtained: kpen ) 0.5 d-1; dlam ) 0.5 mm; kph,ref ) 1.2 d-1.
FIGURE 5. Measured and predicted cumulative volatilization of fenpropimorph. A: Well-exposed scenario. B: Poorly exposed scenario. For both scenarios, a diffusion coefficient in air of 0.5 m2 d-1 was used, and the following values for the parameters to be calibrated were obtained: kpen ) 2.0 d-1; dlam ) 2.0 mm; kph,ref ) 0.5 d-1.
For fenpropimorph, two simulations were performed, reflecting the well-exposed scenario (Figure 5A) and the poorly exposed scenario (Figure 5B). The lack of complete radioactivity and mass balances after application of nonlabeled fenpropimorph complicated the calibration of rate coefficients for penetration and phototransformation. Both, the well-exposed and the poorly exposed scenario revealed a slight overestimation of volatilization of fenpropimorph. Due to the systemic activity of fenpropimorph the counteracting penetration was assumed to be the dominant process over the course of the study (about 60% of the net applied dose). In comparison to fenpropimorph, a lower penetration over the course of the study (about 38% of the net applied dose) was predicted for quinoxyfen, which is in full agreement with the classification of quinoxyfen as a surface-mobile fungicide (Figure 6). As quinoxyfen is known to be photodegraded on the wheat leaf surface (28), its comparatively high predicted transformation of 14% of the applied dose appears realistic. In both scenarios, measured volatilization rates were overestimated by the model approach, but the calculated cumulative volatilization at the end of the simulated period showed reasonably good agreement with experimental findings. This agreement is caused by the low volatilization rates calculated for the final period of the experiment, which is due to the disappearance of plant deposit after day 3 (well-exposed scenario; Figure 6A) and day 8 (poorly exposed scenario; Figure 6B), respectively. Even though the adjustment of the parameters to experimental findings allows for a good agreement between predicted and measured volatilization rates for the applied pesticides, the applicability of the new model approach is
limited by a lack of experimental data on rate coefficients for penetration, photodegradation, and wash-off. For a deeper understanding of these processes additional process studies enabling a thorough controlling of environmental conditions within experiments are required. Studies on photodegradation, including pesticide application to plants at various stages of development, will allow for a characterization of the processes at the leaf-atmosphere interface. Furthermore, the boundary-layer thickness was shown to be a crucial parameter affecting calculated volatilization rates. Due to the simultaneous application of parathion-methyl, fenpropimorph, and quinoxyfen the boundary-layer thickness was expected to be equal for the applied compounds. Therefore, the calibrated values of the boundary-layer thickness covering a broad range of values for the applied compounds between 0.4 mm (quinoxyfen) and 2.0 mm (fenpropimorph) illustrate this parameter to be afflicted with a high degree of uncertainty. PELMO Calculations. Application of the improved PELMO version to the wind-tunnel study allowed for the prediction of the relevant plant and soil processes (Figure 7A-C). Plant processes, including volatilization from crops, penetration into the leaves, and photodegradation, were computed by adjustment of the boundary-layer thickness and the rate coefficients used for the well-exposed scenarios. For the calculation of the soil processes, including volatilization, degradation, and root-uptake, the advanced volatilization module included in PELMO was applied to the soil deposit (30.6% of the net applied dose). PELMO allows the input of the volumetric water content of the soil layers measured in several depths at the beginning of the experiment VOL. 38, NO. 10, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2891
FIGURE 6. Measured and predicted cumulative volatilization of quinoxyfen. A: Well-exposed scenario. B: Poorly exposed scenario. For both scenarios, a diffusion coefficient in air of 0.5 m2 d-1 was used, and the following values for the parameters to be calibrated were obtained: kpen ) 0.5 d-1; dlam ) 0.4 mm; kph,ref ) 1.0 d-1. and calculates the time-dependent water content over the course of the simulation (Figure 8). For the top soil layer (1 mm) PELMO computed decreasing soil moisture up to day 7 and calculated an enormous increase after irrigation was given. In comparison to deeper soil layers, drying and remoistening of the soil was calculated to occur much faster in the top layer leading to considerable fluctuations of water content, which are of importance for the volatilization from soil. However, due to the measurement of water content in the top soil layer being beyond the experimenter’s control, the calculation of water content in the top soil layer is subject to uncertainty (32). During the first days of the simulation decreasing volatilization rates from soil for all compounds were calculated, which is clearly attributed to an increased sorption under dry conditions (Figure 8). After irrigation was given (day 7), the approach installed in PELMO for the consideration of the moisture effect on volatilization from soil led to a marked increase of volatilization of the strongly adsorbed compounds fenpropimorph and quinoxyfen under moist conditions (Figure 7B,C). For weakly sorbed parathionmethyl, this effect was lower, and only a slight increase was computed after moistening of the soil (Figure 7A). Furthermore, due to the comparable low sorption coefficient of parathion-methyl (Table 1), root-uptake (≈0.1%) was predicted after irrigation counteracting volatilization from soil, whereas PELMO did not calculate any penetration into the roots for fenpropimorph and quinoxyfen. For fenpropimorph, computations of the plant processes were in reasonably good agreement with the predictions given in Figure 5A, illustrated by a cumulative volatilization from 2892
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 10, 2004
FIGURE 7. PELMO calculation (semilogarithmic plot). A: Parathionmethyl. B: Fenpropimorph. C: Quinoxyfen. the plants of approximately 3.5% of the net applied dose (Figure 7B). The predicted cumulative volatilization from soil was about 2.6% of the net applied dose, corresponding to 8.5% of the estimated soil deposit. In addition to the increase in soil volatilization after irrigation on day 7 and day 8, a slight wash-off of the remaining plant deposit (≈0.2%) was calculated. The experimental setup used in the windtunnel study did not allow for distinguishing between volatilization arising from soil and plant deposits. However, cumulative volatilization of 6.4% of the applied radioactivity measured during a 13-day wind-tunnel study after soil surface application of 14C-fenpropimorph to gleyic cambisol (19)
FIGURE 8. Soil moisture calculated by PELMO over the course of the wind-tunnel study. suggested the predicted volatilization of fenpropimorph to lie in the range of experimental findings. For the low-volatile quinoxyfen, the predicted cumulative volatilization from soil even exceeded volatilization from plants after day 8 (Figure 7C). This is due to the disappearance of plant deposits after day 7 and the subsequent disrupting of volatilization from plants, while volatilization of the remaining soil deposit still continued. Limitations of the recent understanding of volatilization from soil support the need to establish generalized methods to gauge critical factors, especially phase partitioning coefficients, impacting volatilization immediately after application. Future research should give a detailed understanding of the processes counteracting volatilization from plants and should elucidate the contribution of volatilization from soil after application to the soil-crop system, especially under field conditions with substantial soil coverage. The consideration of volatilization from plants by implementation of a boundary-layer concept in future versions of European registration models will be a significant contribution to recent efforts to harmonize registration procedures for pesticides within the European Union.
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).
Literature Cited (1) Eisenreich, S. J.; Looney, B. B.; Thorton J. D. Environ. Sci. Technol. 1981, 15, 30-38. (2) Van den Berg, F.; Kubiak, R.; Benjey, W. G.; Majewski, M. S.; Yates, S. R.; Reeves, G. L.; Smelt, J. H.; Van der Linden, A. M. A. Water, Air, Soil Pollut. 1999, 115, 195-218. (3) Bedos, C.; Cellier, P.; Calvet, R.; Barriuso, E.; Gabrielle, B. Agronomie 2002, 22, 21-33. (4) Barber, J. L.; Thomas, G. O.; Kerstiens, G.; Jones, K. C. Environ. Sci. Technol. 2003, 37, 3838-3844. (5) Riederer, M.; Daiss, N.; Gilbert, N.; Ko¨hle, H. J. Exp. Bot. 2002, 53, 1815-1823. (6) Taylor, A. W.; Spencer, W. F. In Pesticides in the soil environment: processes, impacts, and modeling; Cheng, H. H., Ed.; SSSA: Madison, WI, 1990; pp 213-269. (7) Majewski, M. S.; Glotfelty, D. E.; Paw, U.; Seiber, J. N. Environ. Sci. Technol. 1990, 24, 1490-1497.
(8) Stork, A.; Witte, R.; Fu ¨ hr, F. Environ. Sci. Pollut. Res. 1994, 1, 234-245. (9) Wolters, A.; Kromer, T.; Linnemann, V.; Scha¨ffer, A.; Vereecken, H. Environ. Toxicol. Chem. 2003, 22, 791-797. (10) Woodrow, J. E.; Seiber, J. N. Environ. Sci. Technol. 1997, 31, 523-529. (11) Jury, W. A.; Spencer, W. F.; Farmer, W. J. J. Environ. Qual. 1983, 12, 558-564. (12) Baker, J. M.; Koskinen, W. C.; Dowdy, R. H. J. Environ. Qual. 1996, 25, 169-177. (13) Wang, D.; Yates, S. R.; Gan, J. J. Environ. Qual. 1997, 26, 10721079. (14) Reichman, R.; Wallach, R.; Mahrer, Y. Environ. Sci. Technol. 2000, 24, 1313-1320. (15) Vanclooster, M.; Boesten, J. J. T. I.; Trevisan, M.; Brown, C. D.; Capri, E.; Eklo, O. M.; Gottesbu ¨ ren, B.; Gouy, V.; Van der Linden, A. M. A. Agric. Water Manage. 2000, 44, 1-19. (16) USEPA (U.S. Environmental Protection Agency) 2001, Water Models; EPA 738-R-00-022. http://www.epa.gov/oppefed1/ models/water/index.html. (17) Leistra, M.; Van der Linden, A. M. A.; Boesten, J. J. T. I.; Tiktak, A.; Van den Berg, F. PEARL model for pesticide behaviour and emissions in soil-plant systems; Descriptions of the processes in FOCUS PEARL v 1.1.1, Alterra-rapport 013, Alterra Green World Research, Wageningen, The Netherlands, 2001. (18) Van den Berg, F.; Wolters, A.; Jarvis, N.; Klein, M.; Boesten, J. J. T. I.; Leistra, M.; Linnemann, V.; Smelt, J. H.; Vereecken, H. In Proc. XII Symp. Pestic. Chem. - Pesticide in air, plant, soil & water system; Del Re, A. A. M., Capri, E., Padovani, L., Trevisan, M., Eds.; La Goliardica Pavese: Pavia, Italy, 2003; pp 973-983. (19) Wolters, A.; Linnemann, V.; Herbst, M.; Klein, M.; Scha¨ffer, A.; Vereecken, H. J. Environ. Qual. 2003, 32, 1183-1193. (20) Leistra, M.; Wolters, A. Water, Air, Soil Pollut. 2004, in press. (21) Wolters, A.; Leistra, M.; Linnemann, V.; Smelt, J. H.; Van den Berg, F.; Klein, M.; Jarvis, N.; Boesten, J. J. T. I.; Vereecken, H. In Proc. XII Symp. Pestic. Chem. - Pesticide in air, plant, soil & water system; Del Re, A. A. M., Capri, E., Padovani, L., Trevisan, M., Eds.; La Goliardica Pavese: Pavia, Italy, 2003; pp 985-994. (22) Stork, A.; Witte, R.; Fu ¨ hr, F. Environ. Sci. Technol. 1997, 31, 949-955. (23) Niehaus, R.; Scheulen, B.; Du ¨ rbeck, H. W. Sci. Total Environ. 1990, 9, 163-172. (24) Grain, C. F. In Handbook of chemical property estimation methods. Environmental behavior of organic compounds; Lyman, W. J., Ed.; McGraw-Hill Book Company: New York, 1982; pp 14/1-14/20. (25) Tucker, W. A.; Nelken, L. H. In Handbook of chemical property estimation methods. Environmental behavior of organic compounds; Lyman, W. J., Ed.; McGraw-Hill Book Company: New York, 1982; pp 17/1-17/25. (26) Klein, M. PELMO: Pesticide Leaching Model, User’s Manual Version 2.01; Fraunhofer-Institute for Environmental Chemistry and Ecotoxicology: Schmallenberg, Germany, 1995. (27) Winberry, W. T.; Forehand, L.; Murphy, N. T.; Ceroli, A.; Phinney, B. In Determination of respirable particulate matter in indoor air; USEPA: Research Triangle Park, New York, 1990; Chapter IP-10. (28) Tomlin, C., Ed.; The Pesticide Manual; British Crop Protection Council and The Royal Society of Chemistry; Cambridge, UK, 2000. (29) Stockmaier, M. Ph.D. Thesis, University of Technology Braunschweig, Germany, 1996. (30) Ebing, W.; Frost, M.; Kreuzig, R.; Schuphan, I. Nachrichtenbl. Dtsch. Pflanzenschutzdienstes (Braunschweig) 1995, 47, 5-9. (31) Staimer, N.; Mu ¨ ller, T.; Kubiak, R. Intern. J. Environ. Anal. Chem. 1996, 65, 183-191. (32) Wolters, A.; Klein, M.; Vereecken, H. J. Environ. Qual. 2004, in press.
Received for review September 26, 2003. Revised manuscript received February 26, 2004. Accepted February 27, 2004. ES035061M
VOL. 38, NO. 10, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2893
