Environ. Sci. Technol. 2005, 39, 2968-2975
Measuring and Predicting Environmental Concentrations of Pesticides in Air after Application to Paddy Water Systems FEDERICO FERRARI,* DIMITRIOS G. KARPOUZAS, MARCO TREVISAN, AND ETTORE CAPRI Istituto di Chimica Agraria ed Ambientale, Universita` Cattolica del Sacro Cuore, Via Emilia Parmense 84, 29100 Piacenza, Italy
In this study, the volatilization of five pesticides applied to an artificial flooded paddy field was assessed using the theoretical profile shape (TPS) and the integrate horizontal flux (IHF) techniques. The dataset derived was utilized to improve the volatilization routine of the rice water quality (RICEWQ) model. The masses of pesticides ethoprophos, procymidone, metalaxyl, chlorpyrifos, and chlorpyrifos methyl volatilized from paddy water and their concentrations in paddy water were determined for a period of 6 d after application. The highest and lowest volatilization losses were observed for chlorpyrifos and metalaxyl, respectively, accounting for 3.3% and 0.03% of their initially applied amount. A rapid pesticide dissipation was evident in paddy water during the study period. The RICEWQ model was used to simulate the fate of pesticides in the artificial paddy system. The Kvolat, an empiric coefficient used by the model as an input parameter, was calculated for all pesticides through model calibration. RICEWQ simulated well the fate of pesticides in paddy water. A significant regression correlation between Henry’s law constant (Hk) and Kvolat of the studied compounds was established which could facilitate the parametrization of the model for describing pesticide volatilization.
Introduction Volatilization was recognized more than 20 years ago as an important process for the loss of pesticides from the areas where they are applied (1-3). Consequently, information on the potential volatility of a pesticide is needed to understand its environmental fate and is therefore required by international registration authorities including Europe (4) and the U.S. (5). The rate and extent of pesticide emission after application depends on physicochemical properties of the pesticide, application parameters, local meteorological conditions during and after application, and characteristics of the surface area receiving the treatment (6). A considerable amount of research on this topic has been conducted to elucidate the mechanisms controlling volatilization of pesticides. Most of these studies have mainly investigated the amounts and rates of pesticides volatilized from treated plant and soil surfaces (3, 7, 8). However, little is known so far about the emission flux of pesticides after * Corresponding author phone: +391523599218; +391523599217; e-mail:
[email protected]. 2968
9
fax:
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 9, 2005
direct application to surface water and thus to flooded paddy fields. Seiber and McChersney (9) measured the volatilization rate of pesticides from paddy water and found that volatilization losses of molinate during the first 72 h after application accounted for approximately 10% of the initially applied amount. Similarly, Ross and Sava (10) reported that about 0.6% of thiobencarb and 9% of molinate applied in a rice paddy field had been lost due to volatilization within the first 24 h after application. These are the only two reported studies which attempted to measure volatilization of pesticides from flooded paddies. Therefore, there is an urgent need for studies investigating the contribution of volatilization on pesticide dissipation in the paddy field environment. Currently, several mathematical models could be used to predict the fate of pesticides in different environmental compartments (11). However, none of these models could adequately describe the environmental fate of pesticides applied in flooded paddy fields (12). The rice water quality model (RICEWQ) is at the moment the only mathematical model which has been used for this purpose. This model has been used for regulatory purposes in the U.S. (13) and has also been validated under European rice scenarios (14, 15). RICEWQ considers that volatilization occurs only by the water-dissolved fraction of the pesticide and calculates the mass volatilized using an empiric volatilization coefficient (13). For most uses of this model to date, the volatilization coefficient was set to zero, as relevant input data are lacking and lumped half-life values including volatilization and photodegradation were used for model parametrization (14, 15). Since volatilization has been identified as a significant process determining the fate of certain pesticides in rice paddies, improvement of the existing volatilization routine of the RICEWQ model is indispensable. The main objectives of this study were to determine the volatilization of certain pesticides from flooded paddy fields and to use the measured dataset for improving the volatilization routine of the RICEWQ model.
Materials and Methods Site Details, Weather Data, and Pesticides. The experiment was employed at the university experimental field site (Universita` Cattolica del Sacro Cuore, latitude 45,06 N, longitude 9,72 E) in Piacenza, Italy (Figure 1). The field is situated on a flat alluvial plain in the Po valley. The soil (030 cm) was characterized as silty loam (sand, 17.9%; loam, 47.1%; clay, 35.0%) with an organic matter content of 2.8%, pH of 8, total nitrogen content of 1.4 g/kg, and organic carbon content of 16.4 g/kg. A paddy field site of approximately 280 m2 (17.5 × 16 m) was prepared on Oct 16. An automated climatic station (model AD2, Silidata spa, Modena, Italy) was installed at the center of the field to record the air and soil temperature (10 cm depth), wind speed and direction, precipitation, air humidity, and solar radiation in 10 min time intervals. In paddies, evapotranspiration could be assumed equal to pan evaporation (16). Therefore, the potential evapotranspiration was calculated with the Penman-Monteith approach using the global solar radiation estimate software RadEst 3.00 (17). Five test substances with varying physicochemical properties were selected (Table 1) for the purpose of the study: the insecticides chlorpyrifos (O,O-diethyl O-3,5,6-trichloro-2 pyridylphosphorothioate), chlorpyrifos methyl (O,O-dimethyl O-3,5,6-trichloro-2-pyridylphosphorothioate), and ethoprophos (O-ethyl S,S-dipropylphosphorodithioate) and the fungicides procymidone (N-(3,5-dichlorophenyl)-1,2dimethylcyclopropane-1,2-dicarboximide) and metalaxyl 10.1021/es048342i CCC: $30.25
2005 American Chemical Society Published on Web 03/10/2005
FIGURE 1. Site map of the field experiment: (a, b) geographic location; (c) field scheme.
TABLE 1. Physicochemical and Environmental Fate Parameters of Pesticides Used in the Study
pesticide ethoprophos procymidone metalaxyl chlorpyrifos chlorpyrifos methyl
formulation emulsifiable concentrate flowable concentrate emulsifiable concentrate emulsifiable concentrate emulsifiable concentrate
water solubility (mg L-1)
vapor pressure (mPa)
Henry’s law constanta (Pa m3 mol-1)
molecular mass (g mol-1)
Koc (mL g-1)
Kdb (mL g-1)
Kdeg(sediment)c (d-1)
Kdeg(water)c (d-1)
application rate (kg ha-1)
750
45.6
1.35 × 10-2
242.3
110
1.803
0.02769
0.0989
1.050
4.5
18.0
1.14
284.1
1500
24.6
0.0330
0.0346
1.050
8400
0.75
1.75 × 10-5
279.3
109.4
1.794
0.0239
0.239
0.730
1.4
2.7
6.76 × 10-1
350.6
8498
139.3
0.0554
0.6923
1.170
4
5.6
4.06 × 10-1
322.5
3300
54.1
0.1385
0.0989
0.840
a Henry’s law constants were calculated by dividing vapor pressure by water solubility both derived at 25 °C and multiplied by the molecular weight of the pesticide. b Kd was calculated by Koc using the formula Koc ) Kd/Foc, where Foc is the fraction of organic carbon in the soil. c Kdeg for paddy water and sediment was calculated using the formula Kdeg ) (ln 2)/(HL), where HL is the half-life of the pesticide assuming first-order kinetics.
(methyl N-(methoxyacetyl)-N-(2,6-xylyl)-DL-alaninate). The pesticides included in the study were not necessarily registered for use in rice paddies, but they were selected to cover pesticides with a wide range of vapor pressure and potential volatility but also to provide a broader picture of pesticide volatilization from surface water systems in which flooded paddy fields are included. In addition, selection of rice-applied pesticides would have limited our study to the use of substances with relatively similar potential volatility, with the exception of the herbicide molinate, which is considered volatile (18). The test substances were all applied in the flooded paddy field with 9.5 cm water on Oct 18 using a calibrated pneumatic spray bar with a hand-carried sprayer and a boom of 5 m width. The water level in the paddy was maintained throughout the study period by regular additions of water when needed. The pH of the aqueous phase of the paddy systems was 6.5, suggesting that chemical hydrolysis of the applied compounds was negligible during the experimental period. Since considerable volatilization normally occurs immediately after application, sample collection was initiated 10 min after application. No crop was grown in the experimental paddy during the study due to the rather short study period (6 d). The short duration of the experiment was selected in accordance with previous studies, where volatilization of pesticides from paddy fields was only measured for a period of 3-6 d after treatment because pesticide volatilization occurs mainly during the first few days after application (9, 10). In addition, several of the pesticides used in paddy fields including molinate, pretilachlor, and cinosulfuron are directly applied onto flooded paddy fields as preemergence or early postemergence treatments, and therefore, no crop is present at the time of application. Measurement of Volatilization. Air Sampling. The airsampling system and the equipment adopted have been
described before (3). Briefly, field measurements of volatilization were performed with four air-sampling stations, each consisting of a glass tube sampler of 20 mm diameter containing a plug of polyurethane foam (PUF). These were connected by a Tygon pipe to an air-sampling pump (model 224-PCEX4, SKC Ltd., U.S.), operated at an airflow of 2 L/min. The PUF sampling plugs were positioned at the center of the plot at Zinst height (Zinst is defined in the section on calculation of volatilization below), and the other three heights were selected by application of the integrated horizontal flux (IHF) method. After sampling, the plugs were placed in their sample jars and stored at -22 °C until analysis. Residue Analysis. The extraction procedure of PUF was carried out using a triple extraction with acetone (50 mL) in an ultrasonic bath, followed by filtration through 10 g of anhydrous sodium sulfate. The extract was concentrated with a rotary evaporator and finally evaporated to a volume of 1 mL under a nitrogen stream, which was used for the gas chromatography-mass Spectrometry (GC-MS) analysis. An Agilent model 6890 series gas chromatograph equipped with an Agilent model 5973 MS detector and a fused-silica capillary column DB-17 (30 m × 0.25 mm i.d. and 0.25 µm film thickness) were used for pesticide quantification. The GCMS operating conditions were injector temperature 250 °C, detector temperature 290 °C, initial temperature 70 °C, temperature increases of 6 °C/min up to 240 °C for 3 min and 10 °C/min up to 270 °C, and hold at 270 °C for 3 min. The carrier gas was He at a flow rate of 1 mL/min, and the injection volume was 1 µL. With these conditions the instrumental detection limit, referred to the volume of the extract, was 0.01 mg/L, and good linearity was achieved (R2 > 0.98) for all pesticides. Identification of the pesticides was achieved through (i) comparison between sample and standard retention times and (ii) comparison between sample VOL. 39, NO. 9, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2969
FIGURE 2. (a) Normalized flux of air (ratio of the horizontal to vertical flux) for different heights calculated with the TSIM. Solid and dotted lines represent unstable and stable atmospheric conditions, respectively. The crossing point between the two lines represents Zinst and Ω values. (b) Wind profile measured on two different days, at 12 p.m. (dotted line) and 12 a.m. (solid line). The wind speed has been measured at three different heights from the water surface: 40, 125, and 240 cm. Dots represent measured points (1 h average), while the line represents a regression line. The constant (9.2 ÷ 9.4) in the exponential function of the regression line corresponds to the roughness length Z0 (cm). (c) An example of a gradient of chemicals in air: concentration of chlorpyrifos in air at different heights in the first samples collected after the treatment. and standard mass spectra. Pesticide concentrations were determined by linear regression technique recovery tests. Air concentrations were calculated by considering the volume of sampled air (19). Extraction recoveries from the PUF plugs for all the studied pesticides ranged between 92.3% and 102.6%. Validation of the Air-Sampling Procedure. Static recovery, retention efficiency, and sampling efficiency were determined to validate the efficiency of the sampling procedure. The static recovery was assessed as follows: an aliquot of each pesticide solution in acetone (25 µL of 187 µg/L) was evenly added to the surface of triplicate PUF plugs, and then the extraction was performed. Untreated control samples and samples fortified only with acetone were also included. The retention efficiency was assessed as follows: Triplicate PUF plugs for each pesticide were fortified with pesticide and connected to the sampling pumps for 3 h in the dark, with a 2 L/min airflow (18.5 °C, 53% relative humidity). Control samples included fortified PUF plugs which were stored at low temperature (
