Environ. Sci. Technol. 2006, 40, 96-102
Volatilization of the Pesticides Chlorpyrifos and Fenpropimorph from a Potato Crop M I N Z E L E I S T R A , * ,† J O H A N H . S M E L T , † J. HILBRAND WESTSTRATE,‡ FREDERIK VAN DEN BERG,† AND R E N EÄ A A L D E R I N K † Alterra, Wageningen University and Research Centre, Post Office Box 47, 6700 AA Wageningen, The Netherlands, and TNO Built Environment and Geosciences, Post Office Box 342, 7300 AH Apeldoorn, The Netherlands
Volatilization of pesticides from crops in the field can be an important emission pathway. In a field experiment with characterization of meteorological conditions, the pesticides chlorpyrifos and fenpropimorph were sprayed onto a potato crop, after which concentrations in the air and on/in the plants were measured. Rates of volatilization were estimated with the aerodynamic profile (ADP), energy balance (EB), relaxed eddy accumulation (REA), and plume dispersion (PD) methods. The volatilization rates obtained with the ADP and EB methods were similar, while some rates obtained with the REA and PD methods in the initial period were lower. Cumulative volatilization of chlorpyrifos during daylight hours (ADP and EB methods) was estimated to be about 65% of the dosage. By far the majority of this volatilization occurred in the first few days. Competing processes at the plant surface had a considerable effect on the dissipation of fenpropimorph, so cumulative volatilization during daylight hours was estimated to be only 7% of the dosage. Plant surface residues were higher than would correspond with the volatilization rate, indicating that penetration into the leaves had occurred.
Introduction Volatilization of pesticides sprayed on plants in the field can be an important pathway of emission to the environment (1). In the context of evaluating pesticides, it is important to know source strength and total extent of this emission. Volatilization is followed by transport beyond the field, exposure of organisms including man via the air, deposition on terrestrial and aquatic ecosystems, and long-range transport (the latter dependent on persistence in the atmosphere). Measurement of pesticide volatilization in the field is possible, but the experiments are time-consuming and expensive. Furthermore, volatilization is dependent on factors such as physicochemical properties of the pesticide (especially vapor pressure), mode of application, plant surface properties, and weather conditions. It is thus desirable to develop possibilities for a certain degree of prediction of pesticide volatilization in the field, based on a limited number of laborious field experiments. * Corresponding author phone: +31 317 474344; fax: +31 317 419000; e-mail:
[email protected]. † Alterra, Wageningen University and Research Centre. ‡ TNO Built Environment and Geosciences. 96
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 1, 2006
Various micrometeorological methods can be used for estimating pesticide volatilization, by a combination of measurements and calculations (2). The field experiments are often laborious, because they require many meteorological measurements and pesticide vapor samplings plus analyses. New technical developments, such as fast detection, sampling, and data processing, can be utilized to develop new, possibly less laborious methods. The number of comparisons between methods for estimating pesticide volatilization is limited, especially for the newer methods and for volatilization from plant canopies. In addition to volatilization, competing processes at the plant surface play a part after spraying of a pesticide (3). The pesticide may penetrate into the leaves depending on its properties, its formulation, plant surface characteristics, environmental conditions, etc. Exposure to sunlight may induce photochemical transformation at a rate dependent on various factors. Describing volatilization from plants thus requires that attention be paid to the significance of the competing processes, which affect the amount of pesticide available for volatilization at the plant surface as a function of time. In the present field experiment on the volatilization of chlorpyrifos and fenpropimorph from a potato crop, concentrations of the pesticides in the air above the crop were measured. Micrometeorological conditions above the crop were measured in detail. The rates of volatilization of the pesticides were calculated by four micrometeorological methods: aerodynamic profile (ADP), energy balance (EB), relaxed eddy accumulation (REA), and plume dispersion (PD). Residues on/in the leaves of the crop were measured at various times after spraying. The extent of volatilization of the pesticides, as a fraction of the dosage, was estimated. Literature data were used to estimate the role of competing processes such as plant penetration and phototransformation in pesticide dissipation at the plant surfaces.
Materials and Methods Experimental Field and Spraying of the Pesticides. The trial field was situated on the experimental farm Oostwaardhoeve near Slootdorp in the Wieringermeer Polder (Province of North-Holland), The Netherlands. The soil was a loamy sea deposit (27% clay, 4.9% organic matter, pH-KCl ) 7.4). A potato crop (variety Agria) was grown on ridges 0.75 m apart. The potato plants were on average 0.56 m high, and soil cover by the plants was estimated visually to be on average 80% (range of 70-90%). Leaf area index was measured to be 2.0 m2 m-2 (n ) 9; SD ) 0.4 m2 m-2). The experimental field (Figure 1) was surrounded by other arable fields with crops having heights less than or equal to that of the potato crop. The nearest wind obstacles were a substantial distance from the field: a row of trees 800 m to the west of the field center and some farm buildings with trees 1000 m to the southwest. The insecticide chlorpyrifos and the fungicide fenpropimorph were sprayed as a mixture onto the potato crop on 25 June 2002. The vapor pressure of chlorpyrifos is reported to be 2.7 mPa at 25 °C (4), which corresponds to 1.4 mPa at 20 °C (assuming the molar enthalpy of volatilization in the Arrhenius equation to be 95 kJ mol-1). Kro¨hl et al. (5) measured the vapor pressure of fenpropimorph to be 2.2 mPa at 20 °C. On the basis of these vapor pressures, substantial volatilization from plants can be expected to occur (6). The solubility of chlorpyrifos in water is 1.4 mg L-1 (at 25 °C) and its log Pow for octanol/water partitioning is 4.7 (4). For fenpropimorph, solubility in water is 4.3 mg L-1 (20 °C) and log Pow is 4.1 (4). 10.1021/es051248x CCC: $33.50
2006 American Chemical Society Published on Web 11/24/2005
FIGURE 1. Layout of the experimental field. The trade products Dursban EC (480 g of chlorpyrifos/L) and Corbel EC (750 g of fenpropimorph/L) were added to the water in the spray tank of a tractor-drawn field sprayer and mixed in thoroughly. The spray boom (width 27 m) was provided with Teejet XR11004 nozzles operating at a pressure of 350 kPa (3.5 bar), which sprayed 0.5 m above the crop. Driving speed of the equipment was measured to be 6.1 km h-1 (n ) 6; SD ) 0.1 km h-1). The combination of driving speed and nozzle delivery resulted in a spray liquid volume of 310 L ha-1. On three occasions during spraying, a sample of 1 mL of spray liquid was taken from the spray cones and added to 100 mL of methanol for chemical analysis in the laboratory. The dosages of chlorpyrifos and fenpropimorph were calculated from the product of spray volume and spray concentration to be 679 and 668 g ha-1, respectively. Spraying started at the west side of the field (upwind of the air sampling spot) at 11:30 h and spraying of the upwind part of the field was completed at 12:00 h (taken as zero time). After this, the part of the field downwind of the sampling spot was sprayed. The dosages of the pesticides were checked by measuring their deposition on potato leaves. Freshly collected leaves (15) were fixed on each of eight polystyrene plates placed at crop height at four locations in the field. The leaves were collected within 30 min of spraying, and within 30 min of collection they were extracted with 100 mL of methanol for chemical analysis. The surface area of the 15 leaves on four plates was measured to be 599 cm2 (n ) 4; SD ) 31 cm2) after extraction. Measuring Pesticides in Air. Air samples for measuring the pesticides within the context of the ADP and EB flux calculation methods were taken in the center of the sprayed potato field (Figure 1). Because of this central sampling position, there was a fetch of at least 100 m over the sprayed crop at any wind direction and a fetch of about 180 m in the
predominant westerly wind directions (Supporting Information, Table S2). The ratio of upwind fetch to maximum measuring height above the evaporating surface is recommended to be at least 100:1 to allow measurement of the steady-state pesticide flux in the internal boundary layer above the field (2). The air samples were taken at 1.0, 1.6, and 1.9 m above the soil surface (top of ridges), so up to 1.34 m above the crop surface. The measuring periods for the ADP and EB methods are given in Tables 1 and 2. Air samples for the REA method were taken close to the sonic anemometer placed in the center of the sprayed potato field, at a height of 3 m above the soil surface. The pesticide sampling system consisted of a sample changer containing three sampling heads, a flow meter, and a vacuum pump. Pesticide vapor in updrafts was sampled via one sampling head, while vapor in downdrafts was sampled via the second head. The samples taken within the deadband range of vertical wind velocities (-0.1 to + 0.1 m s-1) via the third head were not analyzed. The sampling of the pesticides for the PD method was carried out at five points, 30-60 m apart on a line (Figure 1) along the downwind edge of the field (in the plume). Sampling height was 1 m above the soil surface. The measuring periods for the REA and PD methods are given in Tables 3 and 4, respectively. Details on the measurements of the pesticides in the air are given in the Supporting Information. Measuring the Pesticides on/in the Leaves. The amounts of chlorpyrifos and fenpropimorph remaining on the leaves and in the leaf surface were measured at 1.2, 7.6, 29, 73, and 145 h after application. The field was divided into four equal rectangles and the samples were obtained by collecting 15 leaves in each rectangle: (a) from the top of the crop and (b) from half-height of the crop. The collected leaves had a horizontal position and were not covered by other leaves. Within 1 h of sampling, each sample of 15 leaves was combined with 100 mL of methanol in a glass jar and shaken four times by hand (1 min) in 1 h. A fraction of the methanol layer was collected for direct chemical analysis by HPLC. Three leaf samples were collected from the field before spraying; they did not contain measurable residue of either chlorpyrifos or fenpropimorph. Chemical Analyses. Chlorpyrifos in the extracts of XAD polystyrene adsorbent (ADP and EB measurements) was analyzed by gas-liquid chromatography (GLC), by use of an electron-capture detector. Fenpropimorph in the same extracts was analyzed by GLC, equipped with a mass-selective (MS) detector. The MS detector in select ion mode detected fenpropimorph at ion mass 128. Chlorpyrifos and fenpropimorph in the extracts of XAD adsorbent (REA and PD methods) were analyzed by GLC with an MS detector, which detected fenpropimorph at mass 128 and chlorpyrifos at mass 199. Chlorpyrifos and fenpropimorph in the methanol solutions resulting from shaking of the plant leaves were analyzed by high-performance liquid chromatography (HPLC), with UV detection at 220 nm. Details on the chemical analyses are given in the Supporting Information. Collection of Meteorological Data. Equipment for the meteorological measurements was installed in the center of the treated potato field (Figure 1). The ADP method required measurement of wind speed and air temperature at different heights. Net radiation, dew points, and soil heat flux had to be measured additionally for the EB method. The sonic anemometer used in the REA method was placed at a height of 3 m above the soil surface in the center of the field. It was used to measure horizontal and vertical wind speed, wind direction, and air temperature. The data from the sonic anemometer were also used to derive plume dispersion parameters, as input for the PD method. Details on the meteorological measurements are given in the Supporting Information. VOL. 40, NO. 1, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
97
TABLE 1. Data Used in the Calculation of the Volatilization Fluxes of Chlorpyrifos and Fenpropimorph from the Potato Canopy with the ADP Methoda date in 2002
time (h:min)
difference in wind speed, ∆u (m s-1)
25 June
12:35-13:56 14:01-15:34 15:42-17:26 19:20-20:44 9:54-11:32 11:42-13:37 15:12-17:00 16:24-18:00 16:26-18:17
0.512 0.558 0.586 0.393 0.741 0.837 0.790 1.110 1.012
26 June 28 June 1 July
difference in air temp (K) 0.535 0.520 0.437 0.077 0.267 0.398 0.420 0.159 -0.043
difference in concn, ∆c (µg m-3) chlorpyrifos fenpropimorph 4.88 5.52 2.14 0.764 0.381 0.167 0.113 0.0081 0.0081
0.784 0.772 0.189 0.021 bd bd bd bd bd
volatilization flux (mg m-2 h-1) chlorpyrifos fenpropimorph 6.72 7.80 2.96 0.605 0.567 0.286 0.187 0.016 0.014
1.08 1.09 0.262 0.017 bd bd bd bd bd
a Differences for the heights of 1.02 and 1.60 m above soil surface. bd, below detectable level. Wind speed increased with height; temperatures (except once) and concentration decreased with height.
Calculation of the Volatilization Fluxes. In the ADP method, the rate of volatilization is taken to be proportional to the difference in both wind speed and pesticide concentration in air, at two heights above the soil surface (1.02 and 1.60 m). So the flux plane is at 1.28 m (geometric mean). The zero plane displacement was set to 2/3 of crop height (7). Correction factors accounted for the variation in stability of the surface air layer. The values of these correction factors were calculated from empirical relationships, with the gradient Richardson number (Ri) as stability parameter. Ri was calculated from the differences in horizontal wind speed and air temperature measured at the two heights. For details on the ADP method see Majewski et al. (2, 8). In the EB method the sensible heat flux density was calculated from measured net radiation, soil heat flux density, and Bowen ratio (BR) between the sensible and latent heat fluxes. The value of BR equals the ratio between air temperature gradient and the gradient in water vapor pressure, multiplied by the thermodynamic psychrometer constant. The coefficient for turbulent transfer of sensible heat was calculated from the quotient of the sensible heat flux and the vertical air temperature gradient. It is assumed that the coefficient for the transfer of sensible heat is the same as that for the transfer of the pesticide. So the rate of volatilization of the pesticide could be calculated from the product of this transfer coefficient and the measured vertical gradient of pesticide concentration in air. Details of this method have been given before (2, 9). In the REA method the sonic anemometer was used as fast-response sensor for vertical wind velocity in the turbulent eddies. A fast-response valve system in a constant air sampling flow is used for pesticide sampling. Depending on the vertical wind direction, the pesticide vapor is sampled from updraft or downdraft eddies, on different adsorbent cartridges. The vertical flux of the pesticide above the field was calculated according to Businger and Oncley (10). The value of coefficient b0 was measured (11) to be 0.56 in a wide range from unstable to near-neutral surface layer conditions. The difference in vapor concentrations measured in the updrafts and downdrafts was increased by taking a deadband of low wind velocities in the range of -0.1 to +0.1 m s-1 (in which the concentrations were not measured). This range of deadband velocities is expected to give a distinct increase in measured concentration difference, without missing a large fraction of the pesticide vapor. Then the coefficient b0 has to be replaced by coefficient b with a value of 0.48 (12). This b value is in good agreement with that obtained using the Ammann and Meixner (11) equation for the effect of deadband velocity. Standard deviation σw is calculated from the detailed measurements of vertical wind velocity with the sonic anemometer. Details on the use of the REA method in pesticide volatilization studies (with application to the soil surface) have been given before (13, 14). 98
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 1, 2006
In the PD method, the horizontal pesticide concentration profile downwind of the treated field (across the wind direction) is calculated with the Gaussian plume model (15, 16). The treated field is modeled as 97 point sources at zero height, to which a steady-state source strength (e.g., for 1 h) is assigned. The input parameters for the plume dispersion model are derived from the data obtained with the sonic anemometer. The coefficients for dispersion of the pesticide σy (horizontal) and σz (vertical) are calculated from the turbulence parameters in the corresponding direction, the wind velocity profile, and the travel time in the plume. The calculated concentration profile is compared with the horizontal profile measured downwind at the receptor points (1 m high) on the line perpendicular to the wind direction. The source strength of pesticide volatilization from the field is obtained by optimizing the fit between the calculated and measured concentrations at the receptor points.
Results Weather Conditions. Weather conditions, especially during the periods of pesticide vapor measurement, are characterized in Supporting Information Table S2. Days 1 and 2 were predominantly sunny, with occasional clouds, no rainfall, moderate wind speeds (predominantly west-southwest), and a maximum temperature of about 20 °C around noon. In the period from day 3 to day 7, the weather was cloudy with periods of rainfall, comparatively high westerly wind speeds, and a maximum temperature of around 15 °C. Pesticide Concentrations and Fluxes. The concentrations of chlorpyrifos and fenpropimorph in the air at three heights, measured for the ADP and EB flux calculation methods, are presented in Supporting Information Table S3. The concentrations of chlorpyrifos could be measured throughout the sampling period of 6 days. Its concentration showed the expected decrease with height. The concentration of fenpropimorph in air also decreased with height. The concentrations for fenpropimorph were substantially lower than those for chlorpyrifos. From the morning of 26 June onward (second day), fenpropimorph could no longer be detected. The linear regression between concentration in air and the logarithm of height was calculated with SigmaPlot (17). The R2 values in the range of 0.95-1.00 show that the variation of concentration with logarithm of height can to a large extent be explained by the regression. The differences ∆c in concentrations at 1.02 and 1.60 m (Table 1) were calculated from the regression line. This difference is required in the pesticide flux calculations because temperature and relative humidity were determined for these heights. The data used in the calculations for the ADP method and the resulting volatilization fluxes of chlorpyrifos and fenpropimorph are given in Table 1. The volatilization rate of chlorpyrifos decreased substantially in the afternoon and
TABLE 2. Data Used in Calculating the Volatilization Fluxes of Chlorpyrifos and Fenpropimorph from the Potato Field with the EB Methoda volatilization flux (mg m-2 h-1) date in 2002
time (h:min)
net radiation (W m-2)
soil heat flux (W m-2)
Bowen ratio
chlorpyrifos
fenpropimorph
25 June
12:35-13:56 14:01-15:34 15:42-17:26 19:20-20:44 9:54-11:32 11:42-13:37 15:12-17:00 16:24-18:00 16:26-18:17
542.4 531.9 416.9 43.1 372.5 541.2 440.7 261.3 61.0
48.64 38.45 24.17 -0.25 20.83 47.62 25.25 4.42 -1.35
1.009 1.122 1.191 0.507 1.275 1.058 1.692 1.142 -0.526
6.72 8.23 3.11 0.431 0.837 0.316 0.208 0.021 0.038
1.08 1.15 0.27 0.012 bd bd bd bd bd
26 June 28 June 1 July a
Differences in temperature and concentration are given in Table 3 for the ADP method. bd, below detectable level.
TABLE 3. Data Used in Calculating the Volatilization Fluxes of Chlorpyrifos and Fenpropimorph from the Potato Field by the REA Method concn in air (µg m-3) date in 2002
time (h:min)
std deviation vertical wind speed (m s-1)
25 June
12:00-14:00 14:00-15:30 15:30-17:30 17:30-19:30 19:30-20:50 10:40-14:00
0.44 0.49 0.49 0.42 0.31 0.65
26 June
chlorpyrifos updrafts downdrafts 15.79 13.85 7.51 4.66 3.30 1.00
fenpropimorph updrafts downdrafts
9.14 7.40 4.69 2.80 2.11 0.57
8.10 3.11 0.95 0.39 0.187 0.066
4.57 2.11 0.74 0.15 0.173 0.059
volatilization flux (mg m-2 h-1) chlorpyrifos fenpropimorph 5.04 5.44 2.38 1.33 0.65 0.47
2.67 0.84 0.18 0.18 0.01 0.01
TABLE 4. Distribution of Downwind Concentrations Measured in the Air, Used for Calculating the Volatilization Fluxes of Chlorpyrifos and Fenpropimorph from the Potato Field by the PD Methoda
a
time (h:min)
point 1
point 2
concn in air (µg m-3) point 3
12:30-14:00 14:00-15:40 15:40-17:50 17:50-19:50
0.82 5.59 3.02 1.50
2.54 13.24 6.15 3.43
Chlorpyrifos, 25 June 2002 3.62 12.47 5.60 3.31
12:30-14:00 14:00-15:40 15:40-17:50 17:50-19:50
0.37 0.48 0.188
