| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Civil and Environmental Engineering, University of Maryland, College Park, MD 20742
b Environmental Quality Laboratory, USDA, Beltsville, MD 20705
c Center for Environmental Science, University of Maryland, Horn Point Laboratory, Cambridge, MD 21613
* Corresponding author (mcconnel{at}ba.ars.usda.gov).
Received for publication July 5, 2002.
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Abbreviations: HCH, hexachlorocylcohexane • MDL, method detection limit • PUF, polyurethane foam
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
While reductions in nutrient (nitrogen and phosphorus) loads to the Chesapeake Bay are a major focus, one of the goals of the Chesapeake Bay watershed strategy is to achieve a bay free of toxics by reducing or eliminating all controllable sources (Chesapeake Executive Council, 2000). A major source of pesticide contamination in air and surface water is from agriculture (Majewski and Capel, 1995; Larson et al., 1997). Chemical contaminant loads and impacts from agricultural land use pose a potential pollution source to the Chesapeake Bay. Residues of pesticides have been detected in air, water, and rain of the Chesapeake Bay (Glotfelty et al., 1990; Foster and Lippa, 1996; McConnell et al., 1997; Harman-Fetcho et al., 1999, 2000; Lehotay et al., 1999; Liu et al., 2002). However, a complete understanding of pesticide loadings to the Chesapeake Bay region has not been fully developed.
The Delmarva Peninsula (eastern shore) of the Chesapeake Bay is an area of intense agricultural activity with concentrated corn (Zea mays L.) and soybean [Glycine max (L.) Merr.] production. Maryland counties located on the Delmarva Peninsula received 120 Mg atrazine and 220 Mg metolachlor in 1997, representing 54 and 70% of the total atrazine and metolachlor use, respectively, in Maryland (Maryland Department of Agriculture, 1999) (for chemical names of pesticides examined in this study, see Table 1). The high pesticide use rates represent a large potential source for toxic chemicals in soil, surface waters, and the atmosphere. Atmospheric deposition may represent a significant source of pesticides in some Delmarva Peninsula watersheds. Glotfelty et al. (1990) conducted the most relevant study of atmospheric deposition of pesticides on the Delmarva Peninsula in the Wye River estuary. A major limitation of their study was that bulk atmospheric deposition samplers were used (open funnels) and that the sample collection was conducted 20 years ago.
|
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
|
The sampler was equipped with a 0.2-m2 stainless steel funnel that was covered with a lid until the rain sensor activated the lid to open. A Teflon column (12-cm length x 14-mm diameter), which was attached to the bottom of the funnel, was connected to a Teflon filter head containing a 45-mm-diameter glass fiber filter (GF/F, 0.75-µm nominal pore size; Whatman, Maidstone, UK). The filter was designed to trap the operationally defined particle-phase pesticide residues in the rainwater. Attached to the filter head was a solid-phase extraction (SPE) cartridge containing 200 mg of a hyper cross-linked styrene-divinylbenzene copolymer (Isolute ENV+, 6-mL reservoir) to capture the operationally defined dissolved-phase residues. A peristaltic pump (Model 7518-00; Cole-Parmer, Vernon Hills, IL) was connected to the cartridge to pull rainwater through the filter cartridge assembly at a flow rate of 20 to 50 mL min-1 and the extracted water was collected in a 20-L bottle for volume measurement. The pump was activated while the funnel was open.
After a rain event, the site operator removed the cartridge and filter, and shipped the samples on ice packs via overnight mail to the Environmental Quality Laboratory in Beltsville, MD, for analysis. Each rain event was any rain that fell during a 24-h period from 1000 to 1000 h (±1 h) the following day. Samples of less than 50 mL were not analyzed, and cartridges and filters were discarded by the site operators. Three samples had volumes of <50 mL, so they were not analyzed. Between rain events, the funnel, column, and filter holder assembly were cleaned by wiping with lint-free paper and rinsing with 4 to 6 L of distilled water followed by approximately 0.5 L chromatographic-grade acetone (High Purity Solvent; Burdick & Jackson, Muskegon, MI). The collection and extraction of field blanks involved pouring 4 L distilled water into the precleaned rain sampler once a month and treating it in an identical fashion to field samples.
The solid-phase extraction cartridge was dried on receipt using clean nitrogen gas. The absorbed analytes were eluted and concentrated using the same method as described for surface water samples (Liu et al., 2002).
Air
Air samples were collected over a 24-h period once a week from 18 Apr. to 19 Dec. 2000 using a high-volume sampler (Model GPNY 1123; Thermo Andersen, Franklin, MA) at the same site where rain samples were collected. No air samples were collected from 8 August to 18 September as the sampler was not working. Air was pulled at a flow rate ranging from 0.29 to 0.54 m3 min-1 through a 20.3- x 25.4-cm rectangular glass fiber filter (Gelman Type A/E; Pall Corp., Ann Arbor, MI) followed by two cylindrical 7.6-cm-diameter x 7.6-cm-length polyurethane foam (PUF) plugs held within a glass sleeve. From April to August the flow rate was approximately 0.54 m3 min-1 providing sample volumes of 720 to 1050 m3. In September a new motor was installed on the sampler and the motor speed was set at a slightly lower flow rate of approximately 0.3 m3 min-1 leading to lower sample volumes ranging from 420 to 430 m3. The PUF plugs were precleaned using tap water, distilled water, and Soxhlet extraction with pesticide-grade acetone (12 h) followed by ethylacetate (12 h) (Burdick & Jackson High Purity Solvent). Filters were precleaned by baking at 400°C (4 h) and were individually wrapped with aluminum foil before use.
At the end of each sampling period the air filter was folded with particles inside and placed back into the clean foil pocket, and PUF plugs were returned to clean jars. The PUF plugs and filters were kept frozen (-20°C) until extraction. On 14 April, and every Tuesday after 21 November, field blanks of PUF plugs and filters were obtained by pulling air through the air sampler for approximately 1 min and processed in the same way as samples. The PUF plugs were extracted separately in batches of 11 using a Soxhlet extraction apparatus with ethylacetate for 12 h. One clean PUF plug was extracted along with samples for each batch to observe any matrix interference or any contamination from laboratory procedures. Another clean foam plug in each batch was spiked with a mixture of target analytes shown in Table 3 to determine extraction efficiency. All samples were spiked with 25 µL of diazinon–d10 (concentration = 41.4 mg L-1) as an extraction efficiency surrogate. Extracts were reduced to 5 to 10 mL by rotary evaporation and further reduced to 1 mL using a gentle stream of chromatographic-grade (99.9%) N2 gas.
|
Instrumental Analysis
Internal standards of 50 µL anthracene–d10 (30 mg L-1) and chrysene–d10 (74 mg L-1) were added to each sample and each standard just before gas chromatography (GC)–mass spectrometry (MS) analysis. Sample extracts were analyzed for 19 compounds (Table 3). Concentrations and method detection limits of pesticides in the air, surface water, and rain extracts were determined using a Hewlett-Packard (Palo Alto, CA) 5890 gas chromatograph coupled to an HP 5989A mass spectrometer in selected-ion monitoring mode using electron impact (EI) and negative chemical ionization (NCI) mass spectrometry as described by Liu et al. (2002).
Quality Control–Quality Assurance
Pesticide standard solutions prepared from high purity (>99%) neat materials were obtained from AccuStandard (New Haven, CT) in chromatographic-grade methanol. Quantification of each analyte was performed using three to five calibration standards that spanned the concentration range of the samples. The calibration solutions were created by adding varying amounts of standard solutions to laboratory blank extracts. Identification of a particular compound in a sample extract was ensured by a narrow retention time (±0.05 min) and the presence of one of the two qualifying ions within ±20% of the ratio of quantifying ion as found in the matrix standards. Ions monitored have been published elsewhere (Liu et al., 2002).
The method detection limit (MDL) for each compound in each matrix was determined by spiking 5 to 50 ng of each compound (the equivalent to the lowest point of the calibration curve) into 7 to 10 laboratory blanks. For example, for rainwater, eight 4-L distilled water samples were spiked with the standard mixture, poured into the rain funnel, extracted through the same equipment used in the field, and processed in the same manner as described for field samples. Standard deviations of measured replicate concentrations were used to calculate the MDL for each compound according to USEPA standard methods (USEPA, 1986). The MDLs are presented in Table 3. Quantification limits of three times the MDLs were used in this study.
Dissolved-phase rainwater (operationally defined as residues captured on the solid-phase extraction cartridge) collection and extraction efficiency were evaluated by spiking 0.1 µg of the negative chemical ionization and 1.0 µg electron impact mixture compounds into 4-L distilled water and passing through the funnel–filter–cartridge assembly. As shown in Table 3, recovery values were >80% for all target compounds except trifluralin (68%) and pendimethalin (69%). Collection capacity of the solid-phase extraction cartridges was assessed by placing two cartridges in series and extracting several volumes up to 10 L of distilled water fortified with a mixture of target analytes through the sampler. While traces of chlorothalonil, malathion, 
-chlordane, ß-endosulfan, and endosulfan sulfate were observed in extracts from the second cartridge, levels were well below quantification limits. Throughout the study, six laboratory equivalent blanks and eight field equivalent blanks were analyzed. No interfering peaks were found in any of the blank extracts.
On each cruise, a duplicate 10-L river water sample from one selected site, one 10-L river water sample from the same site spiked with 100 to 1000 ng of the target compounds (matrix spike), one 10-L distilled water blank, and one 10-L distilled water spiked with 0.1 to 1.0 µg of the target compounds were extracted together to monitor contamination of the extraction equipment, any matrix interference, and the collection efficiency of the method. The recovery differences were <20% between duplicate samples and therefore average concentrations are reported. Recoveries of spike experiments (n = 5) with distilled water showed that most chemicals were recovered at >70% except trifluralin at 62 ± 6.7% (Table 3). Mean recovery results from river water (n = 8) were not significantly different from distilled water results (analysis of variance [ANOVA] single factor, p 
0.05) except for some compounds with high native levels in the river water (atrazine, CIAT, simazine) (Table 3). Recoveries of cyanazine were lower than in distilled water and recoveries of chlorpyrifos, diazinon–d10, and pendamethalin were higher than in distilled water. These matrix-induced recovery differences may have been less if higher spiking levels had been used. Average surrogate diazinon–d10 recovery for river water samples was 103 ± 12% (n = 40). None of the target chemicals were detected above quantification limits in the laboratory blanks.
Laboratory spike recovery experiments for PUF plugs resulted in average recovery values ranging from 94 to 120% except for trifluralin (38%) and pendimethalin (39%) as shown in Table 3. Mean surrogate recovery value for each sample including spikes and blanks was >86%. No target pesticides were detected in the eight laboratory PUF blanks. In the six field blanks no target chemicals were detected above quantification limits except for metolachlor. Metolachlor was detected in two out of six field blanks just above the quantification limit on 14 Apr. and 21 Nov. 2000. For the vast majority of gaseous air samples, metolachlor concentrations were well above quantification limits and therefore metolachlor data were not adjusted for the possible contribution by the blanks.
The front and back PUF plugs were analyzed separately to assess breakthrough of gas-phase pesticides. Breakthrough to the back PUF plug was generally <30% except for 
-HCH (hexachlorocylcohexane) and -HCH due to their volatile properties. During the warm months and high flow rate (about 0.54 m3 min-1) from 9 May to 1 August, HCHs on the back PUF plug amounted to 50% or more of the front PUF plugs. In the remaining months, breakthrough was much less than 50% although one sample exceeded 60% for -HCH and 100% for -HCH. This may have been due to improper loading of the PUF plugs into the glass sleeve. A similar phenomenon was reported by Jantunen et al. (2000). Results are reported as the sum of both PUF plugs and are likely to be lower limits for HCHs.
Laboratory spike recovery values for target compounds extracted from filter samples were >85%. Mean recovery values and MDLs for air and rain filter samples are presented in Table 3 while those for water filter samples have been reported elsewhere (Liu et al., 2002).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
Pesticide Concentration Data
Air and Rain
Thirty-one air samples and seventy-one rain events were collected during the study and pesticide detection frequency is presented in Table 4. Currently used pesticides like metolachlor, atrazine, simazine, chlorothalonil, endosulfan, and chlorpyrifos were detected in the air and rain with a relatively high frequency. Results from analysis of a representative portion of the rain-filter samples (50% randomly selected) revealed that none of the target analytes were present in the particulate phase at concentrations above quantification limits. Therefore, all rain sample results are operationally defined dissolved-phase concentrations.
Metolachlor and atrazine were detected frequently on the air-filter extracts and trends in particle-phase concentrations follow the same pattern as gas-phase samples. A complete examination and discussion of the vapor particle partitioning behavior of these chemicals will be published in a separate report (Kuang et al., unpublished data, 2003). Since a major focus of this report is air–water gas exchange between the dissolved-phase water and gas-phase air samples, only gas-phase concentrations (defined as those residues captured by the polyurethane foam trap) will be presented here (Table 4).
Metolachlor concentration data in gas-phase air were not normally distributed (P < 0.01) based on four normality tests (Shapiro–Wilk, Kilmogorov–Smirnov, Craner–von Mises, and Anderson–Darling normality tests). Results are better described as a log–normal distribution (P = 0.098–0.14). Metolachlor concentrations in the gas phase ranged from 72 pg m-3 to a maximum value of 10 000 pg m-3 on 9 May 2000 with a median value of 520 pg m-3 (Table 4). Concentrations of metolachlor in air were low in April and then increased to maximal values in May before decreasing in June (Fig. 2) . The maximum concentration in May coincided with the highest expected application use rate and the overlap of corn and soybean planting seasons. Metolachlor was observed in rainfall frequently between April and July with the highest concentration of 1000 ng L-1 on 19 May 2000. The concentrations then declined to below quantification limit (3.6 ng L-1) in late August.
|
Harman-Fetcho et al. (2000) also found metolachlor in the air at the Patuxent River watershed, across the Chesapeake Bay from the Choptank River, during the summer of 1995 with a maximum gas-phase concentration of 2700 pg m-3, three times less than this study. The Patuxent River, however, has only 21% agricultural land use, whereas the Choptank River watershed is a more concentrated agricultural area (57%) (Kutz et al., 2001).
Temporal trends in atrazine gas-phase concentration were similar to metolachlor, but concentrations were one to six times lower, consistent with lower use rates and lower vapor pressure (Table 4). Concentrations were low in April (91–180 pg m-3) and jumped to maximal concentrations in May (230–1600 pg m-3) before declining again in June. The highest atrazine gas-phase concentration was observed on 9 May 2000, corresponding to corn planting activity. Atrazine was only detected from April to June in rainfall with a maximum concentration of 1100 ng L-1 occurring on 11 Apr. 2000. Although the maximum atrazine concentration in rain occurred in April, atrazine concentrations were also elevated in May, coinciding with the high concentrations of atrazine in the air. The maximum atrazine concentration observed on 11 April may be due to the very small volume of rain received (110 mL). Large rain events in May (70–2600 mL) may have diluted atrazine concentrations in rain. These results illustrate the need to view rain results in terms of flux rather than concentration.
In the study by Glotfelty et al. (1990) from the same region, atrazine was also detected in rainfall even in the winter months. In that study, maximum atrazine concentrations of 20 000 pg m-3 in air and 3300 ng L-1 in rain were observed on the Delmarva Peninsula, 2 to 10 times higher than those in our study. The differences may be due to a decline in the use of atrazine since the early 1980s. Majewski et al. (1998) found a maximum concentration of 2800 pg m-3 for gaseous-phase atrazine along the Mississippi River, 1.8 times the maximum value found in our study. Also, concentrations as high as 3000 ng L-1 in rain have been detected at some sites of the U.S. Midwest (Goolsby et al., 1997). The higher concentrations of atrazine in air and rain are probably due to the huge land area dedicated to agriculture in the Midwest.
In contrast to atrazine and metolachlor, simazine was observed in much lower concentrations and less frequently in air. The highest gas-phase concentration was 310 pg m-3, 4 and 30 times lower than that of atrazine and metolachlor, respectively. The very low vapor pressure of simazine and lower application rate (Table 4) may account for the difference. Simazine mirrored the temporal pattern of atrazine in rain except at lower concentrations and for shorter periods.
The other remaining herbicide target analytes such as alachlor and acetochlor were only found in air and rain for short periods during our study. The triazine degradation product CIAT, sometimes called DEA in the scientific literature, was only detected in air in May when atrazine concentrations in air were high. CIAT mirrored atrazine in rainfall with a mean ratio of CIAT to atrazine of 0.28 ± 0.15 (n = 10).
Chlorothalonil is a widely used organochlorine fungicide applied to vegetables, trees, fruits, turf, ornamentals, and other agricultural crops and may be applied on multiple occasions depending on weather conditions (Lehotay et al., 1999). It is highly toxic to fish, aquatic invertebrates, and marine organisms (USEPA, 1999). Chlorothalonil had the second highest maximum concentration in air following metolachlor with a median concentration of 230 pg m-3. Chlorothalonil concentration values in air were low in the spring, increased in the summer, and reached the maximum concentrations of 3500 pg m-3 in mid-September before decreasing later in September (Fig. 2). Air data are missing from August to early September due to problems with sampling equipment. It is possible that the highest air concentrations of chlorothalonil were missed during this period.
Chlorothalonil was also frequently detected in rain ranging from 12 to 2000 ng L-1. Chlorothalonil concentration values were low during the spring and then reached a maximum concentration of 2000 ng L-1 on 14 July. High rain concentrations of chlorothalonil observed in September corresponded to the high concentrations in the air (Fig. 2). The presence of chlorothalonil in rain samples is consistent with its physical properties. The reported Henry's law constant for chlorothalonil is 0.022 Pa m3 mol-1 (Table 4), making it relatively volatile. This compound has also been detected with relatively high frequency in precipitation from the Sierra Nevada of California (McConnell et al., 1998; Lenoir et al., 1999). This indicates that chlorothalonil is persistent in the air and can be transported a significant distance. Harman-Fetcho et al. (2000) also found high concentrations of chlorothalonil in the Patuxent River watershed (maximum = 6800 pg m-3). In a separate study by Lehotay et al. (1999), chlorothalonil was detected in the Patuxent River (2.4–18 ng L-1) at the end of July and in the Choptank River in mid-August.
Endosulfan is an insecticide commonly used on vegetables in the mid-Atlantic region (Lehotay et al., 1999) and is generally applied in a technical mixture containing an to ß isomer ratio of 7:3 (Rice et al., 1997). Endosulfan was detected in all air samples with maximum concentration of 680 pg m-3 of -endosulfan found on 18 July 2000 coinciding with the expected high use rate in summer. Throughout the study, -endosulfan was always observed at higher concentrations than ß-endosulfan, reflecting the composition of the technical mixture and the higher vapor pressure of -endosulfan (Table 4).
In rain, -endosulfan was only observed in the summer with a maximum concentration of 31 ng L-1 on 14 July, which coincided with the highest concentrations in the air. ß-Endosulfan was observed in rain both in the spring and in summer with the highest concentration of 81 ng L-1 also observed on 14 July. Concentrations of ß-endosulfan were approximately a factor of two times higher than those of -endosulfan in the summer, even though approximately half as much ß-endosulfan is present in the technical mixture and less ß-endosulfan was present in the air. Other researchers have found higher concentrations of ß-endosulfan in rain collected near the Great Lakes (Eisenreich et al., 1981; Chan et al., 1994) and in surface water samples in two Chesapeake Bay tributaries (Lehotay et al., 1999) as compared with -endosulfan. This result is expected since ß-endosulfan has a 5.7-fold-higher liquid solubility than -endosulfan (Table 4) and would probably be scavenged more easily by rainfall.
Endosulfan sulfate is a major soil degradation product of - and ß-endosulfan. Endosulfan sulfate, in general, was detected at a lower concentration in the air than - or ß-endosulfan. As with -endosulfan, endosulfan sulfate was only detected for a short period in rain from June to the beginning of August.
Chlorpyrifos is an organophosphate (OP) insecticide that is used for a broad range of lawn and home insecticide products, for agricultural purposes, and for termite treatment (Racke, 1993). The USEPA recently banned chlorpyrifos for home use as a result of a risk assessment associated with the Food Quality Protection Act (FQPA) (USEPA, 2001). The highest concentrations of chlorpyrifos in air were found in May, July, and September with the maximum concentration observed on 23 May at 670 pg m-3. Higher gas-phase concentrations of chlorpyrifos up to 2000 pg m-3 have been detected in the Patuxent River watershed (Harman-Fetcho et al., 2000). The lower chlorpyrifos concentrations in air from our study may be due to less residential land use in Choptank River watershed compared with the Patuxent River watershed. Chlorpyrifos was sporadically detected in rain samples with concentrations in the range of 0.70 to 29 ng L-1.
Diazinon is another widely used organophosphate insecticide applied to control grubs in soil and pests of vegetables, fruit, and tobacco (Lehotay et al., 1999). In contrast to chlorpyrifos, diazinon was detected in air less frequently and at a lower concentration. Concentration values fluctuated during the study period, but tended to increase in the summer and fall with a maximum gas-phase diazinon concentration of 220 pg m-3 on 19 September. The increase in gas-phase diazinon in the summer and fall may be a result of local usage of this insecticide and from increased volatilization rates as air and soil temperatures increase. However, diazinon was only detected once in the rainfall with a concentration of 13 ng L-1 despite its relatively high vapor pressure and low Henry's law constant (Table 4). Diazinon hydrolysis in water may cause concentrations in rainfall to be below quantification limits.
Dissolved-Phase Surface Water
As expected from agricultural practices in the watershed, concentrations of herbicides in water were highest in May, June, and July, subsided in August, and were lowest in December. The highest concentrations during each cruise were observed in the upper (northern) section of the Choptank River at Sites 3 and 4 (Fig. 1). This is reasonable because of heavy agricultural activities in the surrounding counties of Queen Anne's, Talbot, and Caroline. Even though pesticide usage is high in the proximity of the other sampling sites, mixing and dilution of freshwater runoff with saline water from the Chesapeake Bay lowered pesticide concentrations (Table 2) as seen in other studies (Harman-Fetcho et al., 1999) indicating that Choptank River is a source of pesticides to the Chesapeake Bay.
Atrazine and its transformation products CIAT and CEAT (sometimes called desisopropyl atrazine in the scientific literature) were detected in every dissolved-phase surface water sample with maximum values of 3140, 750, and 720 ng L-1, respectively, at Site 3. The temporal and spatial trends in atrazine, CIAT, and CEAT were similar, that is, high concentrations in the spring and summer at Sites 3, 4, and 5. But the ratios of CIAT to atrazine and CEAT to atrazine varied during the year. The site-specific ratio of CIAT to atrazine was 0.12 to 0.88 in May and June, increased to 0.44 to 2.0 in July and 0.84 to 5.0 in August, and eventually reached 1.3 to 6.2 in December. As sites approached the mouth of the river, lower ratio values were observed. A similar trend applied to CEAT to atrazine ratio, only at lower overall values. CEAT and CIAT are major degradation products of atrazine and they are more stable than atrazine (Torrents et al., 1997). Therefore, significant concentrations of CIAT and CEAT may persist in the soil, ground water, and surface water.
Other herbicides such as simazine, metolachlor, cyanazine, alachlor, and acetochlor and insecticides chlorpyrifos, -HCH, and diazinon were also detected in the dissolved-phase and their concentrations are summarized in Table 2. Even though high concentrations were detected in air and rain, chlorothalonil was only detected in August at Sites 5, 6, and 8 with concentrations of 1.0, 1.6, and 0.8 ng L-1, respectively, indicating a short half-life in surface water.
Wet Deposition Fluxes
Relatively high pesticide concentrations in rainwater can be observed during small rain events (Goolsby et al., 1997; Harman-Fetcho et al., 2000), while concentrations may decrease during large rain events due to dilution, despite higher overall pesticide mass. Consequently, to normalize the magnitude of pesticide mass deposited, results of wet deposition here are presented in ng m-2 event-1, called wet flux.
A comparison of the total herbicide, total insecticide, and fungicide wet fluxes throughout the year (Fig. 3) indicates that herbicides contribute the most in atmospheric wet loadings in the spring and early summer, and the fungicide chlorothalonil and insecticides contribute the most in the later months. Compared with herbicides and fungicide, the wet flux of total insecticides was one order of magnitude lower.
|
-HCH) to 81 µg m-2 (chlorothalonil). The total wet flux of simazine in our study compared well with the results obtained by Glotfelty et al. (1990) measured in the early 1980s in the vicinity of the Wye River (a tributary just north of the Choptank), while flux values of atrazine and alachlor measured in our study were reduced due to decreased usage in 2000.
|
Compared with the amounts applied for pest control, the amounts deposited by rain were generally small (Table 5). For instance, the wet load fraction versus usage was <1% for herbicides and <0.05% for organophosphate insecticides (diazinon and chlorpyrifos). However, the total mass of chlorothalonil deposited over the entire watershed during year 2000 is estimated to be about 150 kg, representing 13% of the annual amount of chlorothalonil applied to agricultural land in the watershed. Fourteen percent of the annual amount of -endosulfan applied and 90% of ß-endosulfan were deposited into the watershed. These results indicate that wet deposition is a very important atmospheric process to the overall loading of chlorothalonil and endosulfan to the watershed. The high rain-deposited portions as a percentage of the annual usage for chlorothalonil and endosulfan were quite surprising. These results may suggest a regional transport source from the southern Delmarva Peninsula in Virginia, where large areas of tomato (Lycopersicon esculentum Mill.) and other vegetable crops are grown, and where chlorothalonil and endosulfan are applied quite frequently to prevent and/or control pest diseases (Rice et al., 2001). Unfortunately, no pesticide use information is publicly available for counties in Virginia.
Air–Water Gas Exchange
Air–water gas exchange is an important process for the delivery and removal of semivolatile organic pollutants from natural waters (Bidleman, 1988; McConnell et al., 1993; Hornbuckle et al., 1995). In this report, we have chosen metolachlor to examine in-depth with respect to this process as it was present in air and water throughout the study and information on the temperature dependence of its Henry's law constant is available. To determine the direction and magnitude of air–water gas exchange of pesticides in the Choptank River, the river was divided into seven segments, each of which borders at two neighbor sampling sites as shown in Fig. 1. Measured dissolved-phase pesticide concentrations and water temperatures from the two sites in each segment were averaged. Air samples were only taken at one location and uniform gaseous concentration along the Choptank River was assumed. Surface water pesticide data were paired with air sample data obtained on the same day. In the case where no air sample data are available on the same day as a sampling cruise, average concentrations of two air samples bracketing the sample date were used. Temperature and wind speed data were obtained from a 10-m metrological tower located next to the samplers (Chesapeake Bay Observing System, 2000). Twenty-four-hour-average data were used in the calculation for air–water gas exchange fluxes.
The theoretical basis for calculating the air–water exchange flux of a chemical has been exhaustively discussed in the literature (McConnell et al., 1993, 1997; Hornbuckle et al., 1995; Schwarzenbach et al., 1993, p. 109–123, 182–341). The net flux across the air–water interface is calculated from the mass transfer coefficients across the air-side and water-side boundary layers and the concentration difference between the bulk air and water phases.
The dimensionless Henry's law constant (H') equation corrected for measured water temperatures has been used in a previous study (Harman-Fetcho et al., 2000) and was used here. The mass transfer coefficient KOL is dependent on turbulence due to wind or water currents, temperature, and properties of the chemical such as diffusivity and molecular volume. Details for calculation of overall mass transfer coefficient KOL have been described elsewhere (Schwarzenbach et al., 1993, p. 109–123, 182–341; Bidleman and McConnell, 1995). In this study, KOL values ranged from 1.7 to 2.5 x 10-5 m d-1 when wind speed ranged from 2.6 to 7.3 m s-1.
It is important to consider the level of uncertainty in calculated gas exchange fluxes values from systematic and random measurement errors, systematic errors in the values of H', and uncertainties due to the mass transfer coefficient calculations (Nelson et al., 1998). To assess the random errors in the gas exchange flux calculations, propagation of error analysis was performed. Details of this method have been described elsewhere (Nelson et al., 1998) and are summarized here. Total propagated variance is the linear combination of the weighted contribution of the variances of the mass transfer coefficient, H', and measured concentrations. The term H' is a constant and under any singular condition there is no random error. The error associated with concentration measurements in air and surface water was calculated to be 15 and 12%, respectively. Uncertainty in KOL can be determined by propagating random errors in the air- and water-side transfer velocities. The error associated with KOL was selected at 40% in the study of air–water gas exchange of PAHs in the Chesapeake Bay by Nelson et al. (1998) and was used here. The overall random error of gas-exchange fluxes in this study was 40 to 70% with a majority of the uncertainty arising from KOL except for Segment D in June (overall relative error of 210%) and Segment B in July (overall relative error of 580%), in which no distinctive net air–water gas exchange flux could be observed. For example, the flux of Segment A in May 2000 is -44 ± 19 ng m-2 d-1, with 87% of the error due to uncertainty in KOL, and 13 and <1% of the error due to uncertainty in measured gas-phase and water-phase concentrations, respectively. However, the flux for Segment D in June 2000 is 0.37 ± 0.77 ng m-2 d-1, with 4% of the error associated with KOL, and 54 and 42% of the error associated with measured gas-phase and water-phase concentrations, respectively.
The predicted direction and magnitude of metolachlor flux from air to water is presented in Fig. 4 with calculated error bars. In May, August, and December, the net flux across the air–water interface was dominated by deposition. Higher metolachlor concentrations in the air compared with those in the surface water drove the gas exchange from air to water. The highest deposition flux of 44 ± 19 ng m-2 d-1 was observed in May, when the air concentrations of metolachlor reached their maximum and water concentrations of metolachlor were low. The air sampler was located away from the heavy pesticide usage area and use of concentrations from one site for all segments may underestimate the magnitude of the air-to-water flux.
|
Gas Exchange and Wet Deposition Load to the Choptank River
In May, June, July, August, and December the total metolachlor gas exchange load to the Choptank River was calculated using segment-specific flux rate (Fig. 4) and the surface area of each segment (Fig. 1). The monthly wet deposition load to the river was also calculated. The total dissolved-phase metolachlor mass in the river was calculated based on the volume (Fig. 1) and metolachlor concentration in each segment. The total dissolved-phase metolachlor mass in the river decreased from May (67 kg) to December (9 kg). Except in July when gas exchange acted as a removal process for metolachlor in the river, the input from air–water gas exchange varied from 0.079 to 2.5 g mo-1, representing 0.0048 to 0.12% of the total metolachlor mass in the river. In contrast, the input from the wet deposition in May, June, July, and August was one to three orders of magnitude higher than from gas exchange.
The total mass of metolachlor in December (9 kg) and May (67 kg) can be used as a lower and upper limit of the total mass in the river, respectively. Total wet deposition of metolachlor to the river surface based on measurements at our sampling location was 2.0 kg yr-1. The annual input from wet deposition for metolachlor would be in the range of 3 to 23% of the total mass in the river, indicating that wet deposition of metolachlor was an important source to the river.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  C M Villanueva, G Durand, M-B Coutte, C Chevrier, and S Cordier Atrazine in municipal drinking water and risk of low birth weight, preterm delivery, and small-for-gestational-age status Occup. Environ. Med., June 1, 2005; 62(6): 400 - 405. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF |
