Input Dynamics and Fate in Surface Water of the Herbicide

Jun 28, 2008 - concentration dynamics in the lake tributaries. Concentration-discharge relationships for metolachlor ESA in the main tributary showed ...
0 downloads 0 Views 1MB Size
Download PDF
Environ. Sci. Technol. 2008, 42, 5507–5513

Input Dynamics and Fate in Surface Water of the Herbicide Metolachlor and of its Highly Mobile Transformation Product Metolachlor ESA S E B A S T I A N H U N T S C H A , †,‡ HEINZ SINGER,† SILVIO CANONICA,† ´ P. SCHWARZENBACH,‡ AND RENE K A T H R I N F E N N E R * ,†,‡ Eawag, Swiss Federal Institute for Aquatic Science and Technology, CH-8600 Du ¨ bendorf, Switzerland, and Institute of Biogeochemistry and Pollutant Dynamics (IBP), ETH Zurich, CH-8092 Zurich, Switzerland

Received February 8, 2008. Revised manuscript received April 30, 2008. Accepted May 1, 2008.

A large number of herbicide transformation products has been detected in surface waters and groundwaters of agricultural areas, often even in higher concentrations and more frequently than their parent compounds. However, their input dynamics and fate in surface waters are still rather poorly understood. This study compares the aquatic fate, concentration levels, and dynamicsofthetransformationproductmetolachlorethanesulfonic acid (metolachlor ESA) and its parent compound metolachlor, an often-used corn herbicide. To this end, laboratory photolysis studies were combined with highly temporally resolved concentration measurements and lake mass balance modeling in the study area of Lake Greifensee (Switzerland). It is found that the two compounds show distinctly different concentration dynamics in the lake tributaries. Concentration-discharge relationships for metolachlor ESA in the main tributary showed a high baseflow concentration and increasing discharge dependence during harvest season, whereas baseflow concentrations of metolachlor were negligible and the discharge dependence was restricted to the period immediately following application. From this it was estimated that 70% of the yearly load of metolachlor ESA to the lake was due to groundwater recharge, whereas, for metolachlor, the bigger part of the load, 50-80%, stemmed from event-driven runoff. Lake mass balance modeling showed that the input dynamics of metolachlor and metolachlor ESA are reflected in their concentration dynamics in the lake’s epilimnion and that both compounds show a similar fate in the epilimnion of Lake Greifensee during the summer months with half-lives on the order of 100-200 days, attributable to photolysis and another loss process of similar magnitude, potentially biodegradation. The behavior of metolachlor ESA can likely be generalized to other persistent and highly mobile transformation products. In the future, this distinctly different * Corresponding author phone: +41-44-823 50 85; fax: +41-44823 53 11; e-mail: [email protected]. † Eawag, Swiss Federal Institute for Aquatic Science and Technology. ‡ Institute of Biogeochemistry and Pollutant Dynamics (IBP), ETH Zurich. 10.1021/es800395c CCC: $40.75

Published on Web 06/28/2008

 2008 American Chemical Society

behavior of mobile pesticide transformation products should find a more appropriate reflection in exposure models used in chemical risk assessment and in pesticide risk management.

Introduction Thanks to the seminal work of the U.S. Geological Survey (USGS), it is now well established that persistent transformation products of many herbicides are detected frequently and in high concentrations in streams (e.g., refs 1–4), groundwater (e.g., ref 5) and even precipitation (e.g., ref 6) across the United States. In Europe, the presence of such transformation products in water resources is researched less systematically and accounts are sporadic (e.g., refs 7–10). Nevertheless, all these studies testify to the need of quantifying and predicting the environmental exposure to both parent pesticides and their persistent transformation products to fully assess the risk associated with pesticide use (11). While transformation products that exhibit a similar sorption behavior and reactivity as their parent compound exhibit similar concentrations levels and dynamics in surface water, as is, e.g., the case for atrazine and its transformation product desethylatrazine (DEA) (10), highly polar transformation products, such as the ethanesulfonic acid (ESA) and oxanilic acid (OXA) transformation products of the chloroacetanilide family, differ considerably from their parents in that respect (for structures see Supporting Information (SI) Figure S1). Several studies have confirmed year-round elevated concentrations of ESA and OXA transformation products in both streams and tile-drain samples (3, 4, 12), whereas the parent pesticides showed strong seasonal patterns with a concentration pulse during the application season and very low, often nondetectable concentrations during the remainder of the year. These differences in dynamic behavior are reflected in increasing transformation product to parent compound ratios from the application season toward harvest and winter season (1, 2). The authors of different field studies (1–4, 12) came to the same conclusion that these observations can be explained by the fact that both ESA and OXA transformation products, due to their higher persistence and low retention factor in soils, are efficiently transported to lower soil layers and eventually to groundwater. Consequently, during low-flow and winter conditions, when groundwater recharge makes up an increasing portion of the flow of many streams, this leads to significant, relatively constant concentrations of ESA and OXA in the streams. Although the basic mechanisms responsible for the dynamic behavior of these highly mobile transformation products thus seem to be qualitatively understood, there have been, until recently, no studies that have tried to quantify the contribution of individual fate and transport processes leading to the observed loads and concentrations. Such an understanding is essential, however, to develop better models to predict the concentrations of pesticides and pesticide transformation products in surface waters. The aim of this field study, therefore, was to gain a more detailed, quantitative understanding of what causes the different concentrations dynamics in surface water of a highly mobile transformation product and its parent pesticide. To this end, we combined temporally highly resolved and precise concentration measurements of metolachlor and metolachlor ESA in the main tributaries to Lake Greifensee (Switzerland) with monthly measured lake profiles, mass balance modeling, and laboratory photolysis experiments. Metolachlor OXA concentrations were also measured but will not be discussed VOL. 42, NO. 15, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5507

because the concentrations are about a factor of 5-10 lower than ESA concentrations and therefore often close to the limit of quantification. The Greifensee catchment was chosen as study area because it is very well characterized and therefore presents an excellent experimental field system to study in situ processes. In contrast to an earlier study on atrazine and DEA in the same catchment (10), the concentration dynamics in the tributaries to the lake were measured rather than extrapolated in this study, which allows for a clear distinction between the influence of fate processes and that of the input function on the concentration dynamics in the lake. The specific goals of this study, therefore, were to (i) characterize the different dynamics of metolachlor and metolachlor ESA in the lake tributaries, and thereby evaluate the contribution of groundwater-influenced baseflow versus event-driven water to their overall loads to the lake, and (ii) identify and quantify their loss processes in surface waters. Ultimately, the aim of this study is to contribute to improving the modeling, monitoring, and management of mobile pesticide transformation products.

Materials and Methods Catchment Description and Sampling Scheme. The study area encompassed the catchment of Lake Greifensee, a small eutrophic lake (surface area 8.46 km2; maximum depth 32 m) with regular deep mixing in winter (December to March), located 10 km east of Zu ¨ rich, Switzerland (47°21′N 8°41′E). The hydraulics and morphology of the lake have been described in detail elsewhere (13). The catchment has an area of about 160 km2 and a population of about 100,000. Land use is predominantly agricultural, making up 54% of the total catchment area. Of these, the major crops are corn (10%) and wheat (8%), whereas the crops relevant for metolachlor usage, i.e., soybeans and sugar beet, jointly cover only ∼1% of the agricultural area. Two main tributaries to the lake, Aa Mo¨nchaltorf and Aa Uster, contribute roughly two-thirds of the total discharge to the lake. While the subcatchment of Aa Mo¨nchaltorf is mainly agricultural, the subcatchment of Aa Uster includes the two largest towns in the area as well as Lake Pfa¨ffikon, the outflow of which is flow-regulated. A number of smaller creeks, two on the northeast shore and seven on the southwest shore of the lake, make up most of the remainder of the catchment (see SI Figure S2). The sampling scheme for this study encompassed intensive sampling of the two main tributaries, grab samples for most smaller creeks, and seven monthly lake profiles in 2006. For Aa Mo¨nchaltorf and Aa Uster flow-proportional daily composite samples were collected at two official water authority sampling stations for the period of May 15 to August 31. Additional daily samples were collected during a whole week each in April, September, October, and December 2006, as well as in January and March 2007. The sampling stations were set up such that samples were taken after a constructed drop in the profile, which has been shown to lead to wellmixed conditions for most relevant flow conditions. Daily samples were combined to weekly composite samples for analysis, but daily samples from specific flow events (peak discharge >3 m3/s) were also analyzed separately (see SI Figure S3 for an overview of sampling periods and discharge). Grab samples for the three larger creeks were taken on four occasions in February and March 2007, covering a range of conditions from baseflow to high flow events. Grab samples from the remaining 6 small creeks on the southwest shore were taken on March 2 during a strong flow event. Vertical lake profiles were collected above the deepest point of the lake at six to seven depths in March, monthly from May to September, and in December 2006, with a stainless steel sampling bottle (Friedinger, Lucerne, Swit5508

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 15, 2008

zerland). Vertical concentration profiles at this location were assumed to be representative of the whole lake due to fast horizontal as opposed to slow vertical mixing (13). Lake and river samples were kept in 1 L and 250 mL glass bottles respectively, and stored in the dark at 4° until analysis. Analytical Procedure. All samples were analyzed with online solid-phase extraction prior to liquid chromatography and detection by tandem mass spectrometry. The procedure and its validation are described in the SI and in more detail in ref (14). One major improvement over the method described in ref (14) was the synthesis of an isotope-labeled internal standard for the quantification of metolachlor ESA. d6-metolachlor ESA was synthesized from d6-metolachlor according to the procedure described in ref (15). Using the deuterated internal standards for both compounds, recoveries of analyte standards spiked to samples in concentrations of 100 ng/L were 117 ( 13% and 99 ( 4% for metolachlor ESA and metolachlor, respectively. Limits of detection (LOD) and limits of quantification (LOQ) were 1 ng/L and 5 ng/L for metolachlor ESA, and 0.5 ng/L and 2 ng/L for metolachlor. The precision of the method was determined to be 4 ng/L (95% confidence interval) for metolachlor ESA and 1 ng/L for metolachlor based on repeated measurements of one lake sample and two samples from the main tributaries, both several times within one series as well as across series. The high analytical precision was essential to evaluate the differences in the dynamic behavior of metolachlor and metolachlor ESA and to quantify the fate processes of the two compounds in the lake. Lake Model. Vertical concentration profiles, and hence the mass balance in the lake, were simulated using the computer software AQUASIM (16), which accounts for the morphology of the lake and includes mathematical models for the lake’s hydraulics, chemical fate, and vertical compound transport in the lake. Lake Greifensee is represented by 128 boxes of 25 cm thickness each. Complete horizontal homogeneity is assumed for each layer. Vertical mixing is described by turbulent diffusion with time- and depthdependent diffusivities calibrated from measured vertical temperature profiles. Chemical input into the lake was assumed to stem from inflow through the tributaries exclusively since it is known that subsurface water exchange can be neglected (85% of the depletion rates. The second-order reaction rate constants with OH radical were determined (all at pH 7.0), and the following, very similar values were obtained (all error ranges indicate 95% confidence intervals): (6.1 ( 0.6) × 109 M-1s-1 for metolachlor, (5.6 ( 0.6) × 109 M-1s-1 for metolachlor OXA, and (7.1 ( 0.6) × 109 M-1s-1 for metolachlor ESA. Mass Balance Modeling and Quantification of Elimination Processes in the Lake’s Epilimnion. Based on the results from the laboratory photolysis experiments, indirect photolysis was included in the model as indicated in the method section, i.e., as a function of precursor (NO-3, NO-2) and scavenger (DOC, HCO3-, CO32-) concentrations. Standard deviations of the simulated concentration profiles were obtained by simple linearized error propagation of the standard deviations of all input parameters through the model. For metolachlor ESA, this resulted in reasonably good agreement between modeled and measured concentrations profiles except for the profiles in August and September where modeled concentrations in the epilimnion were too high by about 5-10 ng/L (Figure 3, red line). Taking into account the uncertainty intervals, the August profile still shows a significant difference between measured and modeled concentrations. A closer inspection of model uncertainties reveals the OH radical concentration, and hence the concentrations of its precursors, NO-3 and NO-2, as highly influential. In the model, daily concentration values of these compounds in the first 0.5 m of the lake’s epilimnion are extrapolated from monthly concentration measurements, which vary by less than 4% among the summer months (May-September). In the absence of data on a possible short-term, i.e., diurnal, variability of these species, we assume that our error estimates of (20% for NO-3 and NO-2 concentrations are reasonable and that, as a consequence, photolysis alone is not sufficient as a loss process to explain the low metolachlor ESA epiliminion concentrations in summer. Hence, an additional temperature-dependent first-order loss process was added to the model and its parametrization was derived from optimizing the agreement between mea5512

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 15, 2008

sured and modeled concentration profiles. The implementation of the additional loss process as an Arrhenius-type equation resulted in a final model that fits all observed concentration profiles well. According to the final model indirect photolysis and the additional loss process, potentially biodegradation, contribute about equally to degradative losses of metolachlor ESA in the epilimnion of Lake Greifensee. The half-lives for the two processes during the summer months lie in the range of 160-300 d for indirect photolysis and 60-150 d for the other loss process. Similarly, the results of the earlier study on metolachlor (26) showed that indirect photolysis and another loss process also contributed about equally to the loss of metolachlor from the lake. Finally, the magnitude of these degradative losses was compared to overall disappearance of the two compounds from the lake over the entire simulation period. It was found that the two degradation processes taken together account for the loss of about one-third of both metolachlor and metolachlor ESA from the lake (180 and 1330 g, respectively), whereas the remaining two-thirds (340 and 2730 g, respectively) are lost through the outflow of Lake Greifensee. The two compounds thus exhibit a very similar long-term fate in the lake. Both are quite persistent with similar degradation half-lives that lie in the range of the hydraulic residence time of the lake in summer. Consequences for Transformation Product Assessment and Management. Thanks to the highly precise measurements of metolachlor ESA dynamics in Lake Greifensee and its tributaries, as well as laboratory data allowing for the quantification of direct and indirect photolysis, our study succeeded in establishing the magnitude of the main loss processes of metolachlor ESA in the epilimnion of Lake Greifensse. They were identified as indirect photolysis and another loss process, potentially biodegradation, being similarly important with epilimnion half-lives during the summer months of 160-300 d and 60-150 d, respectively. These results further established that the main factor causing the different concentration dynamics of the parent herbicide metolachlor and its transformation product metolachlor ESA in surface water bodies are not their degradation behavior in those systems, which is quite similar for the two compounds. Rather their different persistence and vertical mobility in soil lead to distinctly different input dynamics from the unsaturated and saturated soil zones into surface water. These findings can likely be generalized to other highly mobile and fairly persistent transformation products of herbicides such as, for instance, the transformation products of benzonitrile herbicides (e.g., dichlobenil, bromoxynil): They can be expected to (i) be present in soil for an extended period after application, causing later runoff events to still exhibit high transformation products concentrations, (ii) to show considerable losses to shallow groundwater through leaching, and hence (iii) to be present in the baseflow component of surface water systems in fairly large concentrations. Our analysis of load contributions showed that these behavioral characteristics are likely to lead to the major part of the total load of mobile transformation products reaching surface water being due to recharge of groundwater into surface streams. This might also lead to mobile transformation products being detected in streams for long periods of time, even after use of the parent compounds has been terminated (23). These findings have major consequences for the assessment and management of mobile transformation products. First, most exposure assessment models for pesticides and veterinary medicines, i.e., for chemicals primarily emitted to land, nowadays include transformation products. However, they do not include recharge of groundwater into surface waters as one possible pathway of exposure to transformation products. Hence, there is a need to modify the models

accordingly to correctly represent the fate of mobile transformation products. This will require the collection of further field data over a broader range of catchments and parent compound/transformation product combinations to be able to calibrate the model processes sufficiently well. Second, in light of our findings, promising options for the management of parent pesticides such as management of critical source areas or the introduction of buffer strips seem of limited efficacy for highly mobile and persistent transformation products. Unless ways are found to enhance their degradation in the soil, reduced input of parent pesticides currently seems to be the only possibility to reduce transformation product losses to water bodies.

Acknowledgments We thank Werner Angst and Heidi Graf for help with the ¨ rich, synthesis of d6-metolachlor ESA, the Labor Veritas, Zu for collection of the samples from the tributaries, Alfred Lu ¨ ck for collecting the lake samples, and Christian Stamm and Paul Capel for reviewing an earlier version of the manuscript.

Supporting Information Available Additional information on sampling, analytical methods, and photolysis experiments; additional measured concentration data. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Battaglin, W. A.; Kolpin, D. W.; Scribner, E. A.; Kuivila, K. M.; Sandstrom, M. W. Glyphosate, other herbicides, and transformation products in Midwestern streams, 2002. J. Am. Water Resour. Assoc. 2005, 41, 323–332. (2) Battaglin, W. A.; Thurman, E. M.; Kalkhoff, S. J.; Porter, S. D. Herbicides and transformation products in surface waters of the Midwestern United States. J. Am. Water Resour. Assoc. 2003, 39, 743–756. (3) Kalkhoff, S. J.; Lee, K. E.; Porter, S. D.; Terrio, P. J.; Thurman, E. M. Herbicides and herbicide degradation products in upper Midwest agricultural streams during August base-flow conditions. J. Environ. Qual. 2003, 32, 1025–1035. (4) Rebich, R. A.; Coupe, R. H.; Thurman, E. M. Herbicide concentrations in the Mississippi River Basin - the importance of chloroacetanilide herbicide degradates. Sci. Total Environ. 2004, 321, 189–199. (5) Kolpin, D. W.; Schnoebelen, D. J.; Thurman, E. M. Degradates provide insight to spatial and temporal trends of herbicides in ground water. Ground Water 2004, 42, 601–608. (6) Goolsby, D. A.; Thurman, E. M.; Pomes, M. L.; Meyer, M. T.; Battaglin, W. A. Herbicides and their metabolites in rainfall: Origin, transport, and deposition patterns across the midwestern and northeastern United States, 1990-1991. Environ. Sci. Technol. 1997, 31, 1325–1333. (7) Hanke, I.; Singer, H.; McArdell, C. S.; Brennwald, M.; Traber, D.; Muralt, R.; Herold, T.; Oechslin, R.; Kipfer, R. Arzneimittel und Pestizide im Grundwasser. gwa 2007, 3, 187–196. (8) Lapworth, D. J.; Gooddy, D. C. Source and persistence of pesticides in a semi-confined chalk aquifer of southeast England. Environ. Pollut. 2006, 144, 1031–1044.

(9) Steen, R. J. C. A.; Van Hattum, B.; Brinkman, U. A. T. A study on the behaviour of pesticides and their transformation products in the Scheldt estuary using liquid chromatography-electrospray tandem mass spectrometry. J. Environ. Monit. 2000, 2, 597– 602. (10) Mu ¨ ller, S. R.; Berg, M.; Ulrich, M. M.; Schwarzenbach, R. P. Atrazine and its primary metabolites in Swiss lakes: Input characteristics and long-term behavior in the water column. Environ. Sci. Technol. 1997, 31, 2104–2113. (11) Gilliom, R. J. Pesticides in U.S. streams and groundwater. Environ. Sci. Technol. 2007, 41, 3409–3414. (12) Phillips, P. J.; Wall, G. R.; Thurman, E. M.; Eckhardt, D. A.; Vanhoesen, J. Metolachlor and its metabolites in tile drain and stream runoff in the Canajoharie Creek watershed. Environ. Sci. Technol. 1999, 33, 3531–3537. (13) Ulrich, M. M.; Mu ¨ ller, S. R.; Singer, H. P.; Imboden, D. M.; Schwarzenbach, R. P. Input and dynamic behavior of the organic pollutants tetrachloroethene, atrazine, and NTA in a lake: A study combining mathematical-modeling and field-measurements. Environ. Sci. Technol. 1994, 28, 1674–1685. (14) Stoob, K.; Singer, H. P.; Go¨tz, C. W.; Ruff, M.; Mu ¨ ller, S. R. Fully automated online solid phase extraction coupled directly to liquid chromatography-tandem mass spectrometry - Quantification of sulfonamide antibiotics, neutral and acidic pesticides at low concentrations in surface waters. J. Chromatogr. A 2005, 1097, 138–147. (15) Aga, D. S.; Thurman, E. M.; Yockel, M. E.; Zimmerman, L. R.; Williams, T. D. Identification of a new sulfonic acid metabolite of metolachlor in soil. Environ. Sci. Technol. 1996, 30, 592–597. (16) Reichert, P. Aquasim - a tool for simulation and data-analysis of aquatic systems. Water Sci. Technol. 1994, 30, 21–30. (17) Canonica, S.; Meunier, L.; von Gunten, U. Phototransformation of selected pharmaceuticals during UV treatment of drinking water. Water Res. 2008, 42, 121–128. (18) Huber, M. M.; Canonica, S.; Park, G. Y.; Von Gunten, U. Oxidation of pharmaceuticals during ozonation and advanced oxidation processes. Environ. Sci. Technol. 2003, 37, 1016–1024. (19) Gerecke, A. C.; Canonica, S.; Mu ¨ ller, S. R.; Scha¨rer, M.; Schwarzenbach, R. P. Quantification of dissolved natural organic matter (DOM) mediated phototransformation of phenylurea herbicides in lakes. Environ. Sci. Technol. 2001, 35, 3915–3923. (20) Zepp, R. G.; Cline, D. M. Rates of direct photolysis in aquatic environment. Environ. Sci. Technol. 1977, 11, 359–366. (21) Aga, D. S.; Thurman, E. M. Formation and transport of the sulfonic acid metabolites of alachlor and metolachlor in soil. Environ. Sci. Technol. 2001, 35, 2455–2460. (22) Krutz, L. J.; Senseman, S. A.; McInnes, K. J.; Hoffman, D. W.; Tierney, D. P. Adsorption and desorption of metolachlor and metolachlor metabolites in vegetated filter strip and cultivated soil. J. Environ. Qual. 2004, 33, 939–945. (23) Kolpin, D. W.; Thurman, E. M.; Linhart, S. M. Occurrence of cyanazine compounds in groundwater: Degradates more prevalent than the parent compound. Environ. Sci. Technol. 2001, 35, 1217–1222. (24) Squillace, P. J. Observed and simulated movement of bankstorage water. Ground Water 1996, 34, 121–134. (25) McDonnell, J. J. A rationale for old water discharge through macropores in a steep, humid catchment. Water Resour. Res. 1990, 26, 2821–2832. ¨ komassnahmen (26) Singer, H.; Anfang, H.-G. Evaluation der O Bereich Pflanzenschutzmittel; Swiss Federal Institute of Aquatic Science and Technology (Eawag), 2005.

ES800395C

VOL. 42, NO. 15, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

5513