Gaseous Deposition Contributes to the Contamination of Surface

Article pubs.acs.org/est

Gaseous Deposition Contributes to the Contamination of Surface Waters by Pesticides Close to Treated Fields. A Process-Based Model Study Carole Bedos,* Benjamin Loubet, and Enrique Barriuso INRA-AgroParisTech, Environment and Arable Crops Research Unit, F-78850 Thiverval-Grignon, France

Downloaded via CALIFORNIA INST OF TECHNOLOGY on June 26, 2018 at 19:36:33 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: The contribution of atmospheric pathways to surface waters contamination by pesticides has been demonstrated. At the local scale, modeling approaches as well as measurements show situations where the contribution of gaseous dry deposition is of the same order or even higher than the drift contribution. The approach presented here consists in estimating the gaseous emissions of pesticides applied in the field, their atmospheric dispersion, and finally their gaseous deposition into aquatic ecosystems at the local scale by running process-based models, that is, the one-dimensional model for pesticide volatilization following application on bare soil (Volt’Air) and the local-scale dispersion and deposition model (FIDES-2D), adapted for pesticides. A significant number of scenarios describes contrasted situations in terms of pedoclimatic conditions (covering 9 years of meteorological data), periods of pesticide application per year, physicochemical properties of the pesticides, and spatial configurations. The identification of the main factors governing gaseous deposition led to the definition of an effective emission factor which explains a large part of the deposition variability. Based on the model outputs, deposition curves are proposed, as a base for a new tool to assess the contribution of gaseous deposition to nontarget ecosystem contamination.

■

INTRODUCTION

According to this data set, the 90th percentile drift deposition after application to field crops decreases from 2.7% of the applied dose at 1 m of the treated field down to 0.29% at 10 m, representing yet 96.6% of the cumulated deposition calculated over 50 m. Models have also been developed (e.g., ref 9). However, gaseous deposition has much less been studied. Siebers et al. (2003)10 measured the gaseous deposition and the drift of lindane, parathion, and pirimicarb into water bodies. The contribution of gaseous deposition to the total deposition was found significant, especially at 50 m from the edge of the field. Quantitatively, up to 0.24% and 0.06% of the applied dose for lindane and parathion, respectively, were found to be deposited at 50 m from the edge of the treated field during the 21 first hours after application. The increased contribution of gaseous deposition with distance was also found when studying

Contamination of surface waters (lakes, ponds, streams, and rivers) by pesticides in agricultural areas is acknowledged.1 The contribution of atmospheric deposition to such contamination has been demonstrated.2−6 Atmospheric deposition is defined as the entry path of airborne substances to aquatic or terrestrial compartments,7 and it comprises wet and dry deposition: wet deposition is the deposition of pesticides dissolved in precipitation while dry deposition is the deposition of pesticides either incorporated in aerosols or in gaseous form. Dry deposition is a process involving surface capture of aerosols and gaseous equilibrium, turbulent diffusion, and gravitational settling. In the vicinity of treated fields, dry deposition may involve drift occurring during the application, which is the deposition of spray droplets transported from the nozzles by the wind, and gaseous deposition occurring during and after the application, which is the deposition of the gaseous fraction of pesticides volatilized from the treated field. Experimental data have been extensively produced on spray drift and compiled in drift tables like Rautmann’s ones.8 © 2013 American Chemical Society

Received: Revised: Accepted: Published: 14250

June 13, 2013 November 8, 2013 November 8, 2013 November 8, 2013 dx.doi.org/10.1021/es402592n | Environ. Sci. Technol. 2013, 47, 14250−14257

Environmental Science & Technology

Article

gaseous deposition of lindane and pendimethalin.11 Lindane and pendimethalin deposition at 50 m during 24 h was estimated between 0.11% and 0.19% and between 0.01% and 0.04%, respectively, for a range of wind directions. The deposition was higher closer to the treated field, 0−5 m (between 0.56% and 0.76% and between 0.12% and 0.17%, respectively). Finally, the vapor pressure of pesticides was the factor most correlated with their gaseous deposition, as demonstrated with extensive sets of experiments in a wind tunnel to measure gaseous deposition to water surfaces of up to 10 pesticides applied on crops.12 Following this experimentation, gaseous deposition curves (expressed as a percentage of the applied dose) as a function of the distance from the edge of the treated field were proposed for 3 classes of compounds differentiated by their vapor pressure.12 Few published models are available to describe pesticide gaseous deposition. One is the empirical model EVA2.0, which is build up based on Fent (2004) experimental results12 and is an update of EVA1.1.13 EVA2.0 has been evaluated by comparison against the process-based deposition model PESTDEP,14 in the EFSA evaluation of the FOCUS Air report.7 EVA2.0 has the advantage of requiring few parameters. However, it is fully empirical, and the physicochemical factors driving gaseous deposition are not explicitly taken into account, which limits its extrapolation to other situations. PESTDEP14 is a 2-dimensional steady state K-model (i.e., that uses the eddy diffusivity K to describe the atmospheric diffusion) for stationary meteorological conditions. The dry deposition in the model is described by the resistance approach. The emission by volatilization is estimated with empirical relationships. Another modeling approach found in the literature was the coupling of an emission model, PEARL15 with an atmospheric dispersion and deposition model, either FIDES16 or OPS.17 These modeling approaches suggest the existence of situations where the contribution of gaseous dry deposition is of the same order or even higher than the drift contribution.14,16 Although drift is the main contribution to total deposition to surface waters at the moment of pesticide application, the longer duration of the volatilization process (several days or weeks) explains why time integrated gaseous deposition can be larger than drift. The gravitational settling of spray droplets also leads to a quicker decrease of the drift contribution to total deposition with distance from the treated field. Pesticide gaseous deposition is expected to depend on environmental conditions and the pesticide properties but also on the geometry of source and deposition surface, together with the nature of the surface in between (water, plants, bare soil). An important factor is the source strength, which drives the atmospheric concentration above the deposition surfaces, depending of the physicochemical properties of the pesticide (e.g., Henry’s law constant for soil application), its persistence in the soil or at the leaf surface, the environmental conditions (soil and air temperature, soil water content, and soil organic matter), and the agricultural practices.18,19 These processes and the factors driving them need therefore to be accounted for to correctly model the contribution of pesticide gaseous deposition to the contamination of nontarget ecosystems close to treated fields. In this study, the gaseous deposition of pesticides is estimated by combining a process-based model of pesticide volatilization following application on bare soil, Volt’Air,20 with a short-range dispersion and deposition model, FIDES-2D,21

adapted for pesticide deposition to water surfaces. These models account for the physicochemical factors mentioned above and were previously validated against several data sets.21−24 A set of scenarios were constructed that covered a range of contrasted (i) pedoclimatic conditions corresponding to 3 locations selected in France, 3 periods of application per year, over 9 years of meteorological data, and (ii) pesticides, corresponding to 26 pesticides covering a large range of physicochemical properties. Finally, several configurations of the source-target geometry were considered: a pond and a 1 m width stream, which were located at various distances from the edge of the treated field. The capacity of the buffer zone inbetween the source and the target was also considered: inexistent and unlimited dry deposition. After a description of the models, the relationship between the modeled gaseous deposition and the physicochemical properties of the compounds is analyzed. Then, the effect of pedoclimatic conditions as well as the source-target geometry is discussed. Simulated gaseous deposition is then compared with the deposition due to drift. Finally, deposition curves are proposed to assess the contribution of gaseous deposition to the contamination of surface waters near treated fields, being based on an analysis of the process-based model outputs run under a wide range of conditions. These deposition curves could be used as a base for a new operational tool to complete the pesticide risk assessment.

■

MATERIALS AND METHOD Simulations are based on an off-line coupling of a model of pesticide volatilization from bare soil, Volt’Air,20 with an atmospheric dispersion and dry deposition model, FIDES-2D 21 . The pesticides source strength and the surface boundary layer characteristics (friction velocity u* and Monin−Obukhov length, LMO) are calculated by Volt’Air and used by FIDES-2D to calculate the atmospheric dispersion and the deposition of gaseous pesticides. Hereafter, a brief presentation of both models is given together with a more detailed description of the developments specifically made for dealing with gaseous deposition of pesticides to water. Scenario building and model runs are also described. Pesticide Volatilization Model Volt’Air. Details on the model Volt’Air are given in the Supporting Information. Briefly, a process-based approach is used, and the model is organized in modules that calculate the energy budget, the vertical transfer of energy (using Fourier’s law), water (using Richards equation), and solutes in the soil profile including the diffusion of gaseous compounds in the air-filled pore space. The soil profile is divided into a user-defined number of layers. The transfers of heat and water are not coupled. Pesticide volatilization to the air is calculated by the local advection analytical solution,25 coupled with the physicochemical equilibrium Jury’s model26 describing the pesticide adsorption between the gas/water and water/soil phases, assuming to be instantaneous for an ideal solution. It runs at an hourly timestep over several weeks. The Volt’Air model originally built to simulate ammonia emissions from a field spread with slurry27 has since been extended to mineral nitrogen fertilizer28 and pesticide applications.20 The capacity of Volt’Air to describe the temperature and humidity surface conditions in a bare soil has been evaluated,29 validating the physicochemical formalisms used in the model for pesticide and ammonia volatilization. 14251

dx.doi.org/10.1021/es402592n | Environ. Sci. Technol. 2013, 47, 14250−14257


Article

Atmospheric Dispersion and Dry Deposition Model FIDES-2D. The two-dimensional dispersion model (FIDES2D) is not detailed here but is described elsewhere.21,30 Briefly, the FIDES model couples dispersion model with a surface exchange model accounting for the exchange of scalars between the roughness height z0 and the surface. The dispersion model is the analytical solution of the advection-diffusion proposed by Philip,31 which assumes power law profiles for the wind speed U(z) and the vertical diffusivity Kz (z) and considers homogeneity in the cross-wind direction (2D). The surface exchange is a resistance analogue model, which relates the source S to the concentration at z0, S = −C(z0)/(Rb + Rsurf). The resistances Rb and Rsurf are described below. The dispersion and surface exchange models are coupled by application of the superimposition principle,32 which relates the concentration C(x,z) at a target location to the volatilization flux S(x0) at a source location, with the use of a dispersion coefficient D(x,z,x0) (in s m−1). Over the source area, S is given by Volt’Air but over the target area S(x) depends on C(x,z0). Hence, the coupling consists in inverting a linear system to retrieve C(x,z0) and therefore S(x). The underlying hypothesis in FIDES-2D is that the studied tracer is conservative after being emitted and that the surface is homogeneous in terms of dynamics: homogeneous roughness length z0 and wind speed U. The degradation of the pesticide within the atmospheric compartment and its wash-out were therefore neglected which is a reasonable hypothesis considering the small transport time at the distances considered in this study. For the developed application to the pesticides, dry deposition of gaseous pesticides to water was modeled using the resistance analogy approach33 and applied to pesticides in PESTDEP,14 which defines a “virtual” resistance to describe the physicochemical equilibrium of the pesticides at the water−air interface: R bw = R surf =

Table 1. Soil Properties for Each Location (From Infosol Database) and Rain Amount Cumulated over Each Period (1 for February/March, 2 for May/June, and 3 for August/ SepSoil Properties for Each Location (From Infosol Database, http://www.orleans.inra.fr/les_unites/us_infosol/ bases_de_donnees) and Rain Amount Cumulated over Each Period (1 for February/March, 2 for May/June, and 3 for August/September). Meteorological data from Agroclim database (https://intranet.inra.fr/climatik/do/welcome) Avignon initial water content (g/kg) pH CO (g/kg) bulk density (kg/m3) soil texture class latitude (deg) longitude (deg) meteorological station rain (mm) (1) (2) (3)

KH(Twater) 1/2

( ) 600 Scw

Guipavas

187

165

7 13 1497

7 17 1365

7 27 1455

3−4 silty sandy loam/silt loam 43.91 4.85 INRA no. 84007004 65.4 104 131.4

4 silt loam

4/5 silt loam/ loam 48.45 −4.41 Meteo-France no. 29075001 217.2 142.5 113.9

48.55 7.64 Meteo-France no. 67124001 87.1 128.5 122.5

properties of each site, needed as model inputs, are given in Table 1, and their geographical location shown in SI Figure 1. Three pesticide application dates were considered as representative of French practices: beginning of March, end of May, and end of August. Twenty-six pesticides were selected covering a large range of physicochemical properties (taken from the Sph’Air database35,36 and given in SI Table 1). Four source-target geometry were considered (Figure 1): a 50 m large pond at the edge of the treated field and a 1 m large stream located at three distances from the edge of the treated field (5, 20, or 50 m). Three widths sizes of the treated field were tested in the pond case: 100, 300, and 500 m. In the stream case, the width of the treated field was set to 100 m. Additionally, for the stream configuration located at 5 m, the buffer zone located between the treated field and the stream was either considered to have a zero dry deposition (Rsurf = ∞) or an unlimited dry deposition (Rsurf = 0). Combining these different configurations, 243 simulations per pesticide for the pond configuration and 324 simulations for the stream one were done. Each simulation was run for two months, one month before the application to let the model equilibrate the soil temperature and water content conditions and one month after the pesticide application. This last month was selected for the pesticide dispersion and deposition modeling. The application timing was fixed at 10:00 A.M. on the 30th day of the simulation and the pesticide application rate was fixed to a normalized value at 1 kg ha −1 to better compare the effect of pesticide physicochemical properties. The results thus obtained may be adjusted afterward with the actual application rate to estimate results closer to actual practices. In Volt’Air, 5 soil layers were defined between depths 0, 0.01, 0.02, 0.05, 0.1, and 0.5 m, and the water transfer was calculated following the Clapp and Hornberger (1978)37 water retention model. In each run, the FIDES-2D model output the timeaccumulated pesticide deposition over the 30 d following application (calculated from the time averaged value of the

1 0.066Sca−0.61u *

1.2510−6u(10)1.6

Entzheim

130

(1)

where Rbw (s m−1) is the quasi-laminar resistance associated with molecular diffusion in the air layer just above the water surface, Rsurf (s m−1) is the virtual resistance accounting for the interface between air and water surface, Sca and Scw are the molecular Schmidt numbers in air and water, respectively (dimensionless), u* (m s−1) the friction velocity, KH the Henry’s law constant (dimensionless) at the temperature of water Twater and u(10) (m s−1) the wind speed at 10 m above ground level. When a buffer zone is considered, dry deposition to a vegetated surface was modeled using a prescribed surface resistance Rsurf (fixed to 0 or ∞) and a resistance Rbv from ref 34 (Supporting Information). Scenario Building and Model Runs. Scenarios were built in order to cover a range of pesticides, pedoclimatic conditions, and application dates that are representative of the French agricultural practices. Three French locations with contrasted climatic conditions were selected (Table1): oceanic (Guipavas), continental (Entzheim), and Mediterranean (Avignon). A time series of 9 years was selected and corresponding data were downloaded from the INRA-Agroclim database. The soil 14252



Article

shows also a wide range of variation (SI Figure 4) and is found maximum between 15 and 20 m from the edge of the applied field, which is in agreement with previous results.16 Factors Driving Gaseous Pesticides Deposition. Pesticide deposition is primarily correlated with the concentration in the air and thus with the pesticide volatilization (Figure 2). Since volatilization is linked with a physicochemical equilibrium between the pesticide in water of the soil and in the atmosphere, the good correlation between volatilization and the Henry constant KH is not surprising. Indeed, in a first order approach, the volatilization flux is equal to the gaseous concentration at the soil surface multiplied by a turbulent transfer coefficient. The range of variation of the turbulent transfer coefficient is much smaller than that of KH, which may cover 20 orders of magnitude, which explains the correlation with KH (Figure 2). However, since the pesticide concentration in the water phase is also in equilibrium with the soil surface with a partition coefficient Koc (for substances that adsorb predominantly to organic matter), it is expected that the correlation between volatilization and KH/Koc is better than with KH alone, which is indeed what we observe in Figure 2 and is in agreement with correlation relationships found in the literature.38 Finally, we also have to consider that the pesticide concentration in the soil also decreases due to its degradation. Assuming a first order reaction type with a degradation rate coefficient k (d−1) and integrating this relationship over time allows calculating a time-integrated residue coefficient of the pesticide over a given period t (see Supporting Information for details), we therefore constructed an effective volatilization indicator Efeff(t):

Figure 1. Schematic of the Volt’Air and FIDES-2D models coupling (top) and of the source-target geometry simulated (bottom). The Volt’Air model calculates the pesticide volatilisation rates and the friction velocity u* and Obukhov length LMO, which are used as input data in FIDES-2D. FIDES-2D then calculates the pesticide concentration at 1.5 m height above the applied field and downwind, and the deposition downwind. Three sizes of the applied field are considered (100, 300, and 500 m). Four geometries are considered: a “pond” of 50 m width located next to the applied field, and a stream of 1 m width separated from the applied field by a buffer zone of width 5, 20, and 50 m. Two surface resistances are tested for the buffer zone: no deposition (Rsurf = ∞) and unlimited deposition (Rsurf = 0). Over the water, Rsurf is a function of the Henry’s law constant KH.

Ef eff (t ) =

■

(3)

The very good correlation between the pesticide volatilization and Efeff(30 d) (Figure 2) indicates that this indicator is a very good descriptor of the volatilization process from soils. Fent (2004) found that pesticide deposition to water was better correlated with vapor pressure Pvap,12 which is not the case here (Figure 2). This is, however, explained by the emission pathway under consideration: in our case, the pesticides are volatilized from bare soil whereas in Fent’s study, the pesticides were volatilized from crop surfaces (winter wheat and sugar beet). The emission was not measured during this study but other authors have shown that volatilization from crop foliage is mainly correlated with Pvap.38,39 The exchange at the air/water interface is not the limiting process for most of the pesticides studied here, as shown by the comparison between the median surface resistances Rsurf and the median boundary layer resistances over water (Rbw) (Figure 3a). However, for two pesticides (trifluralin and fenpropidin), Rbw and Rsurf are of similar magnitude. These two pesticides have the highest Henry constant (KH = 4.18 × 10−3 and 4.39 × 10−3, respectively, dimensionless) in the database used here. There are 26 compounds above a KH = 4 × 10−3 within the Sph’Air database (Figure 3b). Hence, for these compounds, the air−surface exchange process may become limiting (under the modeled meteorological conditions). Comparing Ra+Rbw with Rsurf (Ra being the aerodynamic resistance), Jacobs et al. (2007) found such trigger value for KH.17 They distinguished pesticides with lower KH for which deposition into waterbodies is driven by Rb (Rsurf can thus be neglected and the hydrodynamic behavior of the water body ignored) from pesticides with

deposition rates (in ng m−2 s−1) over the simulation period multiplied by the duration of 30 d), cumulated over the distance from the treated field. Deposition was expressed as kilograms of deposited compound (for an application rate of 1 kg ha−1). The averaged concentration in the air at 1.5 m height and 1 m downwind from the edge of the treated field was also calculated. Comparison to Deposition by Drift. To compare the gaseous deposition to the drift one, the drift deposition (given in % of the application rate) was calculated as a function of the downwind distance from the edge of the treated field x (m) following Rautmann et al. (2001)8: drift = 0.027705x−0.9787

KH DT50(1 − e(−ln(2)/DT50t )) Koc

(2)

RESULTS AND DISCUSSION Emission and Deposition Fluxes. Volatilization rates cover a large range of intensity: simulated cumulative losses after 30 d span from less than 0.0001% up to 99% of the pesticide applied amount, as illustrated in SI Figure 2 for Avignon in June 2001 with a width of treated field of 100 m. Fast or slow emission may occur, with a diurnal variation as shown by the emission fluctuations. SI Figure 3 shows cumulated deposition over time for the pond case. Deposition decreases with the distance from the edge of the applied field and spans 7 orders of magnitude depending on the considered pesticides. Time average concentration in the atmosphere 14253



Article

Figure 2. Correlation table graph between pesticide deposition in the 50 m width pond, pesticide volatilization from the applied field (volatilization), air concentration at 1 m downwind from the applied field, and 1.5 m height (Cair), the pesticide Henry’s law constant (KH), its soil sorption coefficient (Koc), its half-life time DT50, its vapor pressure (Pvap), its water solubility (Sw), the ratio of KH/Koc, and the factor Efeff (see text for definition). The width of the applied field was set to 100 m. Each point on the graphs represents one pesticide.

Figure 3. (a) Comparison between the median boundary layer resistance (Rbw) and the median surface resistance (Rsurf) as a function of Efeff (30 days) for the 26 pesticides in the 5m buffer zone configuration. (b) Cumulative distribution function of the occurrence of pesticides within the database of L’Hermitte and Gouzy (2009)35 for a given Efeff (bold line) and Henry’s law constant KH (dashed line). The vertical lines correspond to the trigger values found for Efeff (gray zone) and KH (dashed line) for the various configurations.

higher KH. When Rb is the controlling parameter, they showed that its parametrization could change the calculated cumulative deposition by a factor of 2 to 3. Influence of Pedo-climatic Conditions on Pesticide Deposition to Surface Waters. The emission strength, which was found to be a key factor governing gaseous deposition of pesticides, depends on physicochemical properties of the compounds as well as on pedo-climatic conditions.18,19 Although we have shown that the former are of first order importance to discriminate pesticides, the secondary may significantly change the amount of pesticides deposited. Rain amount varies significantly between each

location, especially between February and March, when rain amount in Guipavas (oceanic) is 4 times as high as in Avignon (Mediterranean) (Table 1). These conditions lead to rather wet and cold soil conditions as simulated by Volt’Air whereas the soil from Avignon seems to be drier (Figure 4 and SI Figure 5). Soil surface temperatures as simulated by Volt’Air are higher in Avignon than in the 2 other locations, with a significant diurnal variation (reaching 40 °C in August and September at its maximum), whereas the diurnal variation is lower in Guipavas. When considering all simulations for the pond configuration with an emission field of 100 m, the coefficient of variation of the cumulated volatilization and cumulated deposition, which is 14254



Article

Figure 4. Soil water content in the 0−0.01 m horizon (g kg−1) and soil surface temperature (°C) as simuilated by Volt’Air: hourly soil water content and temperature. A, E, and G stands for Avignon, Entzheim, and Guipavas and 1, 2, and 3 corresponds to March, May, and August application periods, respectively.

the maximum in Avignon for the second period (May) and for some cases third period (August) (SI Figure 7), which may be linked with the associated temperature, higher temperature favoring volatilization (when no gaseous adsorption to the soil matrix is considered). Influence of the Source-Target Geometry. An increase of the emission field from 100 to 500 m leads to an increase in the deposition quantity by a factor 2, a behavior explained by the larger atmospheric concentration generated over the target area (Figure 5a). The influence of the intensity of the deposition on the buffer zone located between the edge of the treated filed and the stream is also significant (results not shown) which demonstrates that the nature and the size of the buffer zone have an influence on the deposition to surface waters.

a measure of the influence of pedo-climatic conditions on these variables, varied from 1% to 99% and from 14% to 94%, respectively, depending on the pesticide (SI Figure 6). The coefficient of variation for soil surface temperature and the water content of the soil surface as simulated by Volt’Air were estimated to be 37% and 16%, respectively, for all three soils. A factor up to 13 between the minimum and maximum volatilization flux was found for dicamba, which showed the largest coefficient of variation. The lowest coefficient of variation for the pesticide emission was found for trifluralin and fenpropidin as they are the most volatile ones, and thus, pesticides were completely volatilized (or degraded) before the end of the simulation, whatever the conditions (SI Figure 2). The minimum cumulated pesticide volatilization flux was found in Guipavas for the first period (March) for all pesticides and 14255



Article

to a large extent to the total deposition of pesticides (including drift and gas deposition), on nontarget ecosystems in the vicinity of treated fields. Modeling allows building the standard deposition curves for a wide range of pedo-climatic conditions and a range of different Efeff values, which could then be interpolated to evaluate deposition for a given Efeff. These deposition curves provide the base for the development of an operational and easy to use tool to assess the contribution of gaseous deposition to nontarget ecosystem contamination. These deposition curves have to be tested with data sets obtained under realistic conditions (no data could be found coupling volatilization from soil and deposition on water surfaces). Dealing with the emission pathway, activating the adsorption from the gas phase to the soil matrix would allow better describing the volatilization under dry conditions. This development requires to calculate the associated partition coefficient for all pesticides under consideration and to better describe the soil water transfer under dry conditions. Moreover, as we used here standard laboratory DT50, we should correct them for temperature and moisture. Then, volatilization from crops should be implemented. Dealing with the spatial configuration, different natures of buffer zones (trees, etc.) could be studied as well as the effect of non stagnant water to calculate resulting pesticide concentration in water (i.e., by coupling the model with a surface water model).

■

ASSOCIATED CONTENT

■

AUTHOR INFORMATION

S Supporting Information *

Figure 5. Gaseous deposition of pesticides as a function of the effective volatilization indicator Efeff in (a) the pond case with three applied field width and (b) the stream case with three sizes of the buffer zone. In part b, the applied field was 100 m width and no deposition was considered in the buffer zone (Rsurf = ∞). Drift curves have been added, expressed as cumulated drift up to 50 m in part a and at each distance in part b.

Details on models, calculations, and additional figures and table. This material is available free of charge via the Internet at http://pubs.acs.org.

Comparison with Drift. This study has also shown that the contribution of pesticide gaseous deposition to water surface contamination may be higher than the contribution of drift for pesticides with large Efeff (Figure 5). This conclusion is in agreement with previous tests reported by Asman for the EFSA evaluation of the FOCUS air report 7. The contribution of gaseous deposition is even getting larger when the distance downwind from the treated field increases, since drift decreases faster than gaseous deposition with distance (due to particle settling and evaporation). Experimental data confirm this tendency 10 as well as modeling approaches.14,16 Finally, the coupled Volt’Air-FIDES simulations have shown that gaseous deposition is highly correlated with the quantity of pesticides volatilized from the treated field because deposition to water is mostly not limited by air−water exchange for KH lower than a trigger value found to be in the order of 4 × 10−3 under the conditions studied here. For high KH values, processes of exchange at the air/water interface become the limiting processes. The main operational contribution of this work is the definition of a time-integrated residue indicator (Efeff) calculated from pesticide data (KH, Koc, and DT50) easily found in the pesticide databases and related to the deposition curves. Efeff explains most of the variability of pesticide volatilization over bare soil. The size of the treated field and the size and nature of the buffer zone were also found to be major factors influencing deposition to surface waters. These results moreover show that gaseous deposition may contribute

Notes

Corresponding Author

*Phone: 00 33 1 30 81 55 36. Fax: 00 33 1 30 81 55 63. E-mail: [email protected]. The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank the ONEMA agency for founding this study. Meteorological data were provided by the INRA agroclimatic database from the data exchange network between INRA and Meteo-France.

■

REFERENCES

(1) IFEN. Les Pesticides dans les eaux. Sixième Bilan Annuel. Données 2002, 2004 (42), p34. (2) Van Dijk, H. F. G.; Guicherit, R. Atmospheric dispersion of current-use pesticides: A review of the evidence from monitoring studies. Water Air Soil Pollut. 1999, 115, 21−70. (3) Van Pul, W. A. J.; Bidleman, T. F.; Brorström-Lundén, E.; Builtjes, P. J. H.; Dutchak, S.; Duyzer, J. H.; Gryning, S.-E.; Jones, K. C.; van Dijk, H. F. G.; van Jaarsveld, J. H. A. Atmospheric transport and deposition of pesticides: An Assessment of current knowledge. Water Air Soil Pollut. 1999, 115, 245−256. (4) Waite, D. T.; Bailey, P.; Sproull, J. F.; Quiring, D. V.; Chau, D. F.; Bailey, J.; Cessna, A. J. Atmospheric concentrations and dry and wet deposits of some herbicides currently used on the Canadian prairies. Chemosphere 2005, 58, 693−703. (5) McConnell, L. L.; Nelson, E.; Rice, C. P.; Baker, J. E.; Johnson, W. E.; Harman, J. A.; Bialek, K. Chlorpyrifos in the air and surface water of Chesapeake Bay: Predictions of atmospheric deposition fluxes. Environ. Sci. Technol. 1997, 31, 1390−1398.

14256



Article

(6) Raupach, M. R.; Briggs, P. R.; Ahmad, N.; Edge, V. E. Endosulfan transport II: Modellling airborne dispersal and deposition by spray and vapor. J. Environ. Qual. 2001, 30, 729−740. (7) FOCUS Working Group. Pesticides in Air: Considerations for Exposure Assessment; SANCO/10553/2006 Rev 2; European Union: Brussels, 2008; p327. (8) Rautmann, D.;Review Streloke, M.; Winkler, R. New basic drift values in the authorisation procedure for plant protection products. In Workshop on Risk Assessment and Risk Mitigation Measures in the Context of the Authorization of Plant Protection Products (WORMM); Forster, R., Streloke M., Eds.; Mitt. Biol. Bundesanst. LandForstwirtsch: Berlin-Dahlem, 2001; Issue 381. (9) Holterman, H. J.; ven de Zande, J. C.; Porskamp, H. A. J.; Huijsmans, J. F. M. Modelling spray drift from boom sprayers. Comput. Electron. Agric. 1997, 19, 1−22. (10) Siebers, J.; Binner, R.; Wittich, K. P. Investigation on downwind short-range transport of pesticides after application in agricultural crops. Chemosphere 2003, 51, 397−407. (11) Gottesbüren, B.; Hassink, J.; Pestemer, W.; Richter, T.; Schimdt, H.; Stadler, R.; Zangmeister, W. Deposition of volatilized pesticides on non-target areas. In Pesticide in Air, Plant, Soil, and Water System. Proceedings of the XII Symposium on Pesticide Chemistry. La Goliardica Pavese, Pavia; Del Re A. A. M., Capri E., Padovanil, Trevisan M., Eds.; Goliardica Pavese S.R.L.: Pacienza, Italy, 2003; pp 247−256. (12) Fent, G. Short-range Transport and Deposition of Volatilised Pesticides; Shaker Verlag GmbH: Aachen, Germany, 2004. (13) Winkler, R.; Binner, R.; Gottschild, D.; Koch, W.; Siebers, J. Bewertungskonzept zum nahtransport von pflanzenschutzmitteln infolge exposition über den luftpfad (abtrift, verflüchtigung und deposition). Berichte aus der Biologischen Bundesanstalt für Land und Forstwirtschaft 2002, Issue 110. (14) Asman, W.; Jorgensen, A.; Jensen, P. K. Dry Deposition and Spray Drift of Pesticides to Nearby Water Bodies; Pesticide Research nr66 87-7972-946; Danish Environmental Protection Agency: Kobenhavn, 2003; p 171. (15) Leistra, M.; van der Linden, A. M. A.; Boesten, J. J. T. I.; Tiktak, A.; ven den Berg, F. PEARL Model Pesticide Behaviour and Emissions in Soil−Plant Systems; Descriptions of the Processes in FOCUS PEARL v 1.1.1; Alterra report 013, ISSN 0013-936X, RIVM report 711401 009; Alterra RIVM: Wageningen, 2001; p 115. (16) Durand, E. Modélisation de la Volatilisation, du Transport Atmosphérique et du Dépôt des Pesticides à Proximité de Champ Traité. Master 2, Environnement Spécialité: Evaluation et Gestion des Risques Sanitaires liés à l’environnement; ANSES: Maisons-Alfort, 2010; p 55. (17) Jacobs, C. M. J.; Van den Berg, F.; Van Jaarsveld, J. A. Principles and main uncertainties of a methodology to asess gaseous deposition of pesticides onto small water bodies. In XIII Symposium Pesticide Chemistry, Environmental Fate and Ecological Effects, Piacenza, Italy, 2007; pp 270−277. (18) 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. Emission of pesticides into the air. Water Air Soil Pollut. 1999, 115, 195−218. (19) Bedos, C.; Cellier, P.; Calvet, R.; Barriuso, E.; Gabrielle, B. Mass transfer of pesticides into the atmosphere by volatilization from soils and plants: Overview. Agronomie 2002, 22, 21−33. (20) Bedos, C.; Génermont, S.; Le Cadre, E.; Garcia, L.; Barriuso, E.; Cellier, P. Modelling pesticide volatilization after soil application using the mechanistic model Volt’Air. Atmos. Environ. 2009, 43, 3630−3669. (21) Loubet, B.; Milford, C.; Sutton, A. A.; Cellier, P. Investigation of the interaction between sources and sinks of atmospheric ammonia in an upland landscape using a simplified dispersion-exchange model. J. Geophys. Res. 2001, 106, 24183−24195. (22) Loubet, B.; Génermont, S.; Ferrara, R.; Bedos, C.; Decuq, C.; Personne, E.; Fanucci, O.; Duran, B.; Rana, G.; Cellier, P. An inverse model to estimate ammonia emissions from fields. Eur. J. Soil Sci. 2010, 61, 793−805. (23) Hensen, A.; Nemitz, E.; Flynn, M. J.; Blatter, A.; Jones, S. K.; Sorensen, L. L.; Hensen, B.; Pryor, S. C.; Jensen, B.; Otjes, R. P.;

Cobussen, J.; Loubet, B.; Erisman, J.-W.; Gallagher, M. W.; Neftel, A.; Sutton, M. A. Intercomparison of ammonia fluxes obtained using the relaxed eddy accumulation technique. Biogeosciences 2009, 6, 2575− 2588. (24) Bedos, C.; Rousseau-Djabri; Loubet, B.; Durand, B.; Flura, D.; Briand, O.; Barriuso, E. Fungicide volatilization measurements: Inverse modeling, role of vapor pressure, and state of foliar residue. Environ. Sci. Technol. 2010, 44, 2522−2528. (25) Itier, B.; Perrier, A. Présentation analytique de l’advection. I. Advection liée aux variations horizontales de concentration et de température. Ann. Agron. 1976, 27, 111−140. (26) Jury, W. A.; Spencer, W. F.; Farmer, W. J. Behavior assesment model for trace organics in soil: I. Model description. J. Environ. Qual. 1983, 12, 558−564. (27) Génermont, S.; Cellier, P. A mechanistic model for estimating ammonia volatilization from slurry applied to bare soil. Agric. For. Meteorol. 1997, 88, 145−167. (28) Le Cadre, E. Modélisation de la volatilisation d’ammoniac en interaction avec les processus chimiques et biologiques du sol: Le modèle Volt’Air., INAPG, Ecole Doctorale ABIES, 2004. (29) Garcia, L.; Bedos, C.; Génermont, S.; Braud, I.; Cellier, P. Assesssing the ability of mechanistic volatilization models to simulate soil surface conditions. A study with the Volt’Air model. Sci. Total Environ. 2011, 409, 3980−3992. (30) Loubet, B.; Milford, C.; Hensen, A.; Daemmgen, U.; Erisman, J.W.; Cellier, P.; Sutton, M. A. Advection of NH3 over a pasture field, and its effect on gradient flux measurements. Biogeosci. Discuss. 2009, 6, 163−196. (31) Philip, J. R. The theory of local advection: 1. J. Meteorol. 1959, 535−547. (32) Thomson, D. J. Criteria for the selection of stochastic models of particle trajectories in turbulent flows. J. Fluid Mech. 1987, 180, 529− 556. (33) Liss, P. S.; Slater, P. G. Flux of gases across the air−sea interface. Nature 1974, 247, 181−184. (34) Hicks, B. B.; Baldocchi, D. D.; Meyers, T. P.; Hosker, R. P.; Matt, D. R. A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities. Water Air Soil Pollut. 1987, 36, 311−330. (35) L’Hermite, N.; Gouzy, A. Identification and Classification of Preoccupying Pesticides for Health and the Environment: Sph’Air, a Decision Support System for Air Monitoring; No. DRC-09-8027800977A; INERIS: Verneuil-en-Halatte, France, 2009; p 280. (36) L’Hermite, N.; Gouzy, A.; Le Gall, A. C.; Bedos, C.; Bonicelli, B. Identification and classification of pesticides for health and environment: Sph’Air, a decision support system for air monitoring. Abstracts Book of the 5th European Conference on Pesticides and Related Organic Micropollutants in the Environment, Marseille 2008, pp638−641. (37) Clapp, R. B.; Hornberger, G. M. Empirical equations for some soil hydraulic properties. Water Resour. Res. 1978, 14, 601−604. (38) Woodrow, J. E.; Seiber, J. N.; Baker, L. W. Correlation techniques for estimating pesticide volatilization flux and downwind concentrations. Environ. Sci. Technol. 1997, 31, 523−529. (39) Leistra, M. Estimating Input Data for Computations on the Voltilisation of Pesticides from Plant Canopies and Competing Processes; 1256, ISSN 1566-7197; Alterra: Wageningen, 2005; p 80.

14257


Gaseous Deposition Contributes to the Contamination of Surface

Recommend Documents