JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 8 September 2005
Published in J Environ Qual 34:1873-1882 (2005)
DOI: 10.2134/jeq2005.0049
© 2005 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Leu, C.
Right arrow Articles by Stamm, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Leu, C.
Right arrow Articles by Stamm, C.
Agricola
Right arrow Articles by Leu, C.
Right arrow Articles by Stamm, C.
Related Collections
Right arrow Ecological Risk Assessment
Right arrow Organic Compounds
Right arrow Agricultural Pesticides
Right arrow Water Pollution

TECHNICAL REPORTS

Surface Water Quality

Comparison of Atrazine Losses in Three Small Headwater Catchments

Christian Leua,b, Heinz Singera, Stephan R. Müllera,c, René P. Schwarzenbacha and Christian Stamma,*

a Swiss Federal Institute for Environmental Science and Technology (EAWAG) and Swiss Federal Institute of Technology (ETHZ), Ueberlandstrasse 133, CH-8600 Dübendorf
b Current address: Syngenta, Jealott's Hill International Research Centre, Bracknell, RG42 6EY, UK
c Current address: Swiss Agency for the Environment, Forests and Landscape (SAEFL), 3003 Bern-Ittingen

* Corresponding author (christian.stamm{at}eawag.ch)

Received for publication February 10, 2005.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Understanding the processes causing herbicide transport to surface waters is crucial to determine mitigation options to reduce these losses. To this end, we investigated the atrazine (2-chloro-4-ethylamino-6-isopropylamino-1,3,5-triazine) transport in three agricultural catchments (1.1–2.1 km2) in the watershed of Lake "Greifensee" (Switzerland). In 1999, atrazine application data were recorded for all three catchments. Time proportional samples were taken at a high temporal resolution at the catchment outlets. Extremely wet conditions caused large relative losses from the catchments, ranging between 0.6 and 3.5% of the amount applied. Most of the atrazine load was due to event-driven diffuse losses from the fields. Farmyard runoff contributed less but caused the highest concentrations (up to 31 µg L–1) in the brooks. The maximum concentrations due to diffuse losses varied between 1.2 and 8.2 µg L–1 among the catchments. Despite different absolute concentration levels, the concentration time-series were very similar. It seems that the travel-times within the catchments were mainly controlled by the rainfall pattern with little influence of the catchment properties. These properties, however, caused the relative losses to vary by a factor of 6 between the catchments. This variability could be partly explained by differences in the connectivity of the fields to the brooks and by their hydrological soil properties. A comparison of the losses from the three catchments with those from the entire watershed of Lake Greifensee demonstrated that they were representative for the larger area. Hence, the study results provide a good data set to evaluate distributed models predicting herbicide losses.

Abbreviations: ELISA, enzyme linked immunosorbent assay • SPE-GC/MS, solid-phase extraction gas chromatography–mass spectrometry • WWTP, wastewater treatment plant


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE USE OF HERBICIDES is indispensable for conventional intensive farming because it guarantees high production levels and product quality. One important unwanted side effect is the frequent and widespread herbicide transfer from the site of application to surface waters. Such compounds pose a risk to aquatic life, particularly in brooks and rivers draining agricultural areas. Their concentrations may exceed water quality standards for the protection of aquatic organisms as well as for drinking water usage. The ecological effects of herbicides on nontarget organisms depend both on concentration levels and on the duration of the exposure. Hence, to assess the impact of herbicide contamination in running water it is necessary to adopt adequate monitoring strategies reflecting the dynamics of herbicide exposure. To mitigate these contamination problems, a thorough understanding of the processes controlling herbicide losses is needed.

In agricultural areas, surface water contamination with herbicides originates from point and diffuse sources, e.g., farmyards acting as point sources and agricultural fields resulting in diffuse losses (Kreuger, 1998; Müller et al., 2002; Neumann et al., 2002). After improper handling of herbicides on farmyards (e.g., during filling or cleaning of sprayers) herbicides may be lost to surface waters by direct farmyard runoff or through waste water treatment plants (WWTPs). Knowledge of the relative importance of these sources within a given region is important for defining effective mitigation strategies. In the catchment of Lake Greifensee (146 km2) in Switzerland, we found that losses through WWTPs of herbicides used only in agriculture contributed <30% to the total lake input (Gerecke et al., 2002). This result indicates that diffuse losses dominate herbicide input into surface waters under conditions typical for large parts of Swiss agricultural land.

Diffuse herbicide losses show a strong seasonal pattern. The major part of losses from agricultural fields usually occurs during the first significant discharge event after application, leading to peak concentrations as well as to maximum loads during a relatively short period (Donald et al., 1998; Gruessner and Watzin, 1995; Larson et al., 1995; Leu et al., 2004a; Müller et al., 1997; Schottler et al., 1994; Thurman et al., 1991). This implies that herbicides reaching surface waters at high concentrations have rather short residence time in the soil. Hence, fast transport mechanisms like surface runoff or preferential flow into subsurface drainage systems are the dominating processes. The extent of herbicide losses to surface water streams depends on various factors including weather, soil characteristics, herbicide properties, and agricultural management practices (Blanchard and Lerch, 2000; Capel and Larson, 2001; Kladivko et al., 2001; Leonard, 1990; Lerch and Blanchard, 2003; Wauchope, 1978). Weather conditions, in particular the duration, intensity, and timing of rain before and after application, have a strong influence on herbicide losses. Thus, total herbicide loads from a given catchment may vary considerably from year to year (Capel and Larson, 2001; Müller et al., 1997; Richards et al., 1996). Under similar weather conditions, herbicide losses may also vary between different regions due to different runoff potential of the soils. Several studies in watersheds of very different scales (37–20400 km2) show a tendency that the larger the percentage of well-drained soils the lower the herbicides losses to surface waters (Blanchard and Lerch, 2000; Capel and Larson, 2001; Lerch and Blanchard, 2003; Richards et al., 1996). However, results from such large-scale studies are often confounded by influences of other factors like land management practices or smaller-scale topography that can hardly be assessed in sufficient detail at those scales. Further uncertainty is often introduced by the limited quality of the herbicide use data at hand.

Studies at the scale of smaller agricultural catchments—in the order of few hundred hectares—have the potential to investigate the dominating factors in more detail without the limitation of studies focusing only on edge-of-field losses inherent to field- or plot-scale experiments. In recent work, we have demonstrated the usefulness of controlled experiments in such a small catchment (2.1 km2) for obtaining results that may serve as a scientific basis for concrete advice to farmers. We could show that within such a small area, the relative herbicide losses, that is, expressed as percentage of the applied amounts, varied by a factor of 56 (Leu et al., 2004a, 2004b) in neighboring subcatchments after a controlled application. This variability was mainly caused by field characteristics influencing the losses by fast transport. The permeability of the soils, topography, and location of subsurface drainage systems were identified as key factors. Further, the high temporal sampling resolution helped to distinguish between direct farmyard runoff and diffuse losses, with the latter dominating the total load.

These findings suggested that avoiding application to risk areas, specifically areas having a high vulnerability to lose herbicides by fast transport, could substantially reduce herbicide concentrations in surface waters. However, this conclusion was based on results obtained in a single year and from a single small catchment. Furthermore, the study year was unusually dry before and after the herbicide applications. The study presented here aims at verifying the conclusions obtained previously by quantifying the atrazine losses and the concentration dynamics from three small headwater catchments in a different year, which was very wet. With this study, we address (i) the influence of inter-annual weather variability on the loss dynamics and on the percentage of applied amounts lost to the brooks, (ii) the variability with regard to atrazine losses between three catchments, (iii) the relative importance of point vs. diffuse sources, and (iv) the influence of catchment characteristics on concentration time-series at the outlets of the catchments. Furthermore, based on measurements at the scale of the entire Greifensee watershed in the same year, it is possible to test the representativeness of the three headwater catchments for a larger region (Fig. 1) . To achieve these goals, it was necessary (i) to assess in detail the atrazine application patterns in the study areas and (ii) to quantify properly the herbicide export behavior, which required a high temporal sampling resolution at each catchment outlet.



View larger version (61K):
[in this window]
[in a new window]
  		Fig. 1. Map of the headwater Catchments 1 to 3 located within the watershed of Lake Greifensee (G). The different colors indicate the spatial distribution of the hydrological soil groups (red, high risk for topsoil water saturation; orange, moderate risk; yellow, low risk; green, forests). Black dots, cornfields connected to the brook (for details see text); black hatching, confields that are not connected; P, Lake Pfäffikon.

 

   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Catchment Description
The three catchments have surface areas between 1.1 and 2.1 km2 (Table 1) and are located about 25 km southeast of Zurich. Agricultural land with a moderate slope for Switzerland (Table 1) covers 80 to 93% of the respective areas (Peyer et al., 1997). Typical field sizes are about 1 ha. The soils have developed from two different geological substrates. In Catchment 2 and in about half of Catchment 1, a formation called Obere Süsswassermolasse consisting of tertiary deposits is found. The geological underground of the rest of Catchment 1 and of Catchment 3 is a moraine material dated from the Würm Glaciation (Zingg, 1934). Predominant soil types (FAO Taxonomy) in the catchments are (Calcaric and Eutric) Cambisol, Luvisol, Gleysol, and Gleyic Cambisol (Table 1). Proportion of poorly drained gleyic soils is higher in Catchment 1 (38%) than in the other two catchments (15 and 19%). The dominant soil texture is loam. As the three catchments are located within an area of only 10 km2, they are exposed to similar weather conditions. The 30-yr average annual precipitation is 1327 and 1342 mm, as recorded at two rain gauges of the Swiss national weather service (Swiss meteo) located close to the catchments (Fig. 1). About 12, 35, and 17% of Catchments 1 to 3, respectively, are drained with tiles at ~1.4 m and spaced at ~14 m. Inlets connected to streams partially collect surface runoff along roads and from some farmyards.


View this table:
[in this window]
[in a new window]
  		Table 1. Characterization of the catchments.

 
Field Measurements
Two rain gauges (Grüningen, Hinwil) of the Swiss national weather service (Swiss meteo) located close to the catchments recorded daily rain amounts (Fig. 1). Within each of the catchments, an additional rain gauge was installed to record rain amounts every 10 min. At the outlets of the catchments (Fig. 1), water-level was monitored continuously by pressure transducers (PR-46, Keller AG, Switzerland). In addition, water-conductivity (TetraCon 325 conductivity cell and LF 325 conductivity meter, WTW Weilheim, Germany) and temperature (Temperature Probe 107, Campbell Scientific, USA) were recorded at time intervals of 5 min and data were stored in data loggers (CRX 10; Campbell Scientific). The dilution method with NaCl as tracer was used to calibrate a level-discharge relation for each outlet (Herschy, 1995), each based on 16 calibration points over the entire discharge range gauged during the sampling period.

Two portable water samplers (ISCO 3700, ISCO; and Manning 4901, Manning Environmental) sampled each outlet of the catchments. Each sampler was equipped with 24 plastic 1-L bottles. The samplers took discrete, 1-L, time-proportional samples. At each station, one of the samplers was triggered by critical discharge levels and took samples at intervals between 7 and 15 min, the other sampled continuously at intervals between 0.5 and 4 h. During the entire study (April–August 1999), a total of 3740 samples were transferred to glass bottles and transported to the laboratory. The samples were collected within 2 h during the first discharge events. Subsequently, we extended this period to a maximum of 3 d during base flow periods. In the laboratory, the samples were stored in the dark at 4°C until analysis. Repeated analysis of control samples demonstrated that atrazine concentrations were stable until the end of all measurements a few months after collection.

The atrazine application data was obtained by interviewing all farmers producing corn (Zea mays L.) in the three catchments. Since all farmers participated, complete information was obtained including the exact field locations, the timing of application, and the amounts applied.

Analytical Methods
Out of the total of 3740 samples taken in the field, 1370 were selected for analysis to thoroughly characterize the atrazine concentration dynamic in the streams. Atrazine content was determined using an enzyme linked immunosorbent assay (ELISA) (Giersch, 1993). Briefly, well plates (Greiner F-Elisa, Huber & Co.) were coated with an antibody (K4E7) that binds primarily to atrazine. Then, aliquots (200 µL) of the unfiltered samples and a chromogen (atrazine covalently bound to an enzyme) were added. After washing the wells, atrazine concentrations could be determined photometrically, as atrazine and the chromogen bind to the antibody. A six-point standard curve was measured for each analytical run. In 129 of these 1370 samples, atrazine was additionally determined using a solid-phase extraction gas chromatography–mass spectrometry (SPE-GC/MS) method similar to that of Bucheli et al. (1997). Briefly, filtered (0.45-µm pore size) 1-L samples were spiked with isotope labeled d5–atrazine as internal standard and subsequently enriched on Carbopack B (Supelco). Afterward, substances were separated on a HRGC 8000 (Fisons, England) equipped with a 10-m DB 17 MS column (Agilent Technologies). Detection was performed with a MD 800 mass spectrometer (Fisons Instruments) in the positive electron impact mode. Limit of quantification for atrazine was 5 ng L–1. The comparison of the atrazine concentrations in these samples proved the reliability of the ELISA method as atrazine concentrations measured by the two techniques were highly correlated (R2 = 0.95) with only a small bias. Generally, concentrations determined with the ELISA were somewhat smaller compared with the SPE-GC/MS method (median relative deviation of ELISA measurements from SPE-GC/MS quantifications: –11%).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Application
In Switzerland, atrazine usage is restricted to spring (May and June) applications on corn. In 1999, this application period was extremely wet with 300 and 220 mm of precipitation recorded at the Swiss meteo station Grüningen during May and June, respectively. These amounts are substantially larger than the 30-yr average (May, 130 mm; June, 155 mm). Due to these weather conditions, >90% of the total atrazine used in the catchments were applied exceptionally late (between 26 May and 29 June) (Fig. 2A) . A total of 11, 4.2, and 5.4 kg atrazine were applied on 18 cornfields in Catchment 1, on 7 cornfields in Catchment 2, and on 8 cornfields in Catchment 3, respectively. These fields represent 8.2, 5.9, and 4.2%, respectively, of the total areas of Catchments 1 to 3. Application rates on the fields varied between 0.18 and 0.9 kg a.i. ha–1.



View larger version (32K):
[in this window]
[in a new window]
  		Fig. 2. (A) Cumulated atrazine amounts applied within the Catchments 1 to 3; (B) discharge; and (C) atrazine concentrations and cumulative atrazine loads at the outlet of the Catchments 1 to 3. The concentration peaks caused by direct farmyard runoff are marked with arrows. The maximum concentrations caused by farmyard losses varied from 0.7 and 31 µg L–1 (note the different concentration axes).

 
Point vs. Diffuse Sources
In a previous work (Leu et al., 2004a), we demonstrated that herbicide losses from direct farmyard runoff and agricultural fields can be differentiated based on the duration of the concentration peaks, the triggering of farmyard losses by little rainfall and using the concentration ratios between parent compound and metabolites. In Fig. 2—showing the atrazine time-series at the outlets of all three catchments—the farmyard peaks identified based on these criteria are marked by arrows. Even though the losses from the farmyards caused the highest atrazine concentrations (up to 31 µg L–1), they accounted for only 4 and 8% of the total atrazine load from the two Catchments 1 and 3, respectively, during the relevant period from 1 June until 31 July. Farmyard losses could be ruled out in Catchment 2 because no farmyard was located within the area, which was corroborated by the fact that no concentration peak typical for point sources was observed (Fig. 2).

Influence of Weather Conditions on Diffuse Atrazine Losses
In all three catchments, the major atrazine losses occurred during a rather short period. They started with the first significant discharge event after the first application and ended about 20 d after the last application (Fig. 2). This seasonal variation is in line with findings in numerous other investigations (e.g., Ng and Clegg, 1997; Yuan et al., 2000; Donald et al., 1998; Richards and Baker, 1993). The atrazine concentrations increased almost simultaneously with increasing discharge at the catchment outlets with maximal losses occurring during peak flow. This event-driven concentration dynamics confirms that the herbicide was mainly lost by fast transport mechanisms. In an experimentally controlled study (Leu et al., 2004b), we could show that both surface runoff and preferential flow contributed significantly to the losses in Catchment 1, and that they hardly differed with respects to the timing of the concentration peaks. In that study conducted in a different year in Catchment 1 (Leu et al., 2004a), weather conditions were much drier with only 54 mm mo–1 rainfall before the herbicide application and 119 and 270 mm rain during the following 2 mo, respectively. Under these conditions, the first significant discharge event in Catchment 1 did not occur until 23 d after the controlled application and caused 70% of the observed atrazine load. Despite similar runoff volumes, the loads and concentrations were much smaller during successive events. The rainfall pattern and the loss dynamic during the study presented in this paper were different with much more rainfall before and right after the applications. In Catchment 1, we monitored a series of four successive events occurring within 11 d during which there were no further atrazine applications (Fig. 3) . During the first three events, the loads increased slightly from 58 to 68 g (Fig. 3), although peak concentrations declined from 8.2 to 5.2 µg L–1. The concentration decrease was compensated by the increasing maximum discharges induced by similar rain amounts (Fig. 3). This series of events indicates that the losses to surface waters were not only source-controlled (controlled by the amount of herbicide available for fast transport in the thin transport-relevant zone at the soil surface) but also by the volume of fast flow discharged during a given event. The latter increased during this series most probably due to the increasing soil wetness.



View larger version (24K):
[in this window]
[in a new window]
  		Fig. 3. (A) Rain intensity and cumulated rain amount within Catchment 1; (B) discharge and cumulative discharge; and (C) atrazine concentration as well as cumulative atrazine load at the outlet of Catchment 1 during a sequence of four discharge events, occurring within 11 d without applications. The concentration peak caused by farmyard runoff is marked with an arrow.

 
The different weather conditions had a strong effect on the variability in the total amounts lost from Catchment 1 during the 2 study years. In 1999, the losses from Catchment 1 expressed as percentage of the mass applied (relative losses) were 3.5% and four times larger than in the much drier year 2000 (Leu et al., 2004a). This year-to-year variability is well within the range reported by Capel and Larson (2001), who compiled measurements from 21 watersheds and reported a factor of 20 as the maximum inter-annual variability.

Site Characteristics and Atrazine Losses
The differences between the losses from the three catchments were larger than the inter-annual difference discussed above for Catchment 1. Until 1 mo after the last application, 380, 30, and 34 g of atrazine were lost from the three catchments, respectively. These losses correspond to 3.5, 0.7, and 0.6% of the applied herbicide amounts. Due to their spatial proximity and due to similar application dates (Fig. 1 and 2), the weather conditions before and after the atrazine applications were quite similar in all three catchments. Cumulated rain amounts recorded from 1 May to 30 June at Swiss meteo stations Grüningen and Hinwil (Fig. 1) were 512 and 501 mm, respectively. The temporary installed rain gauges within the catchments recorded 461, 412, and 540 mm rain, respectively, for Catchments 1 to 3 during the same period. These recordings are likely to be more error prone than the Swiss meteo measurements. However, their data clearly indicated that all major rain events occurred in all three catchments (data not shown). There is no indication that variable weather conditions or different application dates caused the loss differences between the three catchments.

Cumulated discharge accounted for 161, 114, and 260 mm, respectively from Catchments 1 to 3 during the period of major atrazine losses (1 June–31 July). These values correspond to between 40 (Catchment 2) and 73% (Catchment 3) of the rain in the respective catchments. These discharge ratios are not related to the difference in relative atrazine losses. This is not surprising since they reflect the hydrological properties of the entire catchments, but atrazine was only applied on <8% of each catchment area. Thus, the large variability in losses was mainly caused by characteristics of the individual cornfields. In a more detailed study within Catchment 1 (Leu et al., 2004b), we have shown that connectivity and hydrological soil properties of the treated fields were crucial for explaining the spatial variability. The connectivity relates first to the topographic situation of the cornfields determining whether surface runoff can directly reach the stream and second to the presence of subsurface drainage systems assessed by means of detailed construction maps.

Qualitatively, connectivity and hydrological soil properties can partially explain the variability between the catchments found in this study as well. We estimated that between 60 and 77% of the total atrazine amount was applied on fields that were directly connected via tile drains or surface runoff pathways to the open watercourses. Assuming that atrazine was only lost from these fields, the spatial variability of relative losses decreased slightly with losses ranging from 4.4, 1.1, and 1% for the three catchments, respectively. The decreased variability indicates that connectivity is important to consider but it was not the single dominant factor controlling the variability of the losses.

Generally, the initiation of fast flow processes requires the saturation of the topsoil (e.g., McCoy et al., 1994; Anderson and Burt, 1990) unless Hortonian overland flow occurs, which is not the dominant process in temperate, humid regions like our study areas. Therefore, one may expect that larger atrazine losses occurred from soils prone to water logging as compared with well-drained soils. The previous work (Leu et al., 2004b) supported this expectation based on the Swiss hydrological soil classification (Table 2). We relied our interpretation on this classification based on soil morphological features since no direct measurements of soil hydrological properties were undertaken. In Catchment 1, a much larger fraction of atrazine was applied to connected fields that have a high tendency for water logging (52%) compared with Catchment 2 (5%) (Fig. 4) . This agrees well with the four times larger losses from connected fields in Catchment 1 (4.4%) compared with those in Catchment 2 (1.1%). Based on the same argument, one would, however, expect the losses from the connected fields of Catchment 3 being similar to those from Catchment 1 since their hydrological classifications are similar (Fig. 4). This was clearly not the case, since the value of Catchment 3 (1.0%) was four times smaller than that of Catchment 1 (4.4%).


View this table:
[in this window]
[in a new window]
  		Table 2. Swiss classification of hydrological soil groups. Among the cornfields of Catchments 1 to 3, six (italic) of the nine hydrological soil groups, which have different tendencies for topsoil saturation (low, moderate, high), could be identified as shown in parentheses.

 


View larger version (56K):
[in this window]
[in a new window]
  		Fig. 4. Portions of total atrazine amount applied on cornfields of Catchments 1 to 3 that are connected to surface water by surface runoff and/or tile drain, classified for their tendency to topsoil saturation according to the hydrological soil groups.

 
These results suggests that the three static soil categories (Table 2) are not sufficient to represent the temporally variable differences of soil hydrological conditions and/or that the hydrologic soil properties alone do not explain the entire variability between connected fields. Additional factors may have influenced the extent of losses, for example (i) the slopes of the fields and their micro-topography influencing surface runoff, (ii) their susceptibility to surface sealing, and (iii) the differences in relative contributions of surface runoff and subsurface drainage. Larger losses may be expected if losses occur by surface runoff (Kladivko et al., 2001) or by both mechanisms compared with tile-drain losses alone.

Hydrologic soil properties were found to have a significant influence on the extent of atrazine losses by Blanchard and Lerch (2000) as well as Capel and Larson (2001) since losses from runoff prone soils (Hydrologic soil Groups C and D; USDA, 1972) were larger than from soils with a lower runoff potential (Groups A and B). Both studies, however, compared losses from significantly larger areas than investigated in this study, encompassing a much wider spectrum of soil types. Thus, a more pronounced influence of hydrological characteristics can be expected on variability of losses.

Influence of Site Characteristics on Loss Dynamics
The maximum atrazine concentrations associated with diffuse losses differed by a factor of about six (8.2, 1.4, and 3.7 µg L–1, respectively) between the Catchments Outlets 1 to 3. These concentrations are within the range of maximum levels (1 to ~100 µg L–1) observed in other European and North American surface waters (David et al., 2003; Donald et al., 1998; Garmouma et al., 1998; Gruessner and Watzin, 1995; Hyer et al., 2001; Jaynes et al., 1999; Lerch and Blanchard, 2003; Richards and Baker, 1993; Williams et al., 1995).

Despite the variability in maximum concentrations, the concentrations time series at the catchment outlets showed important similarities. The concentrations increased at all outlets during all major discharge events occurring until 1 mo after the last application (Fig. 2). Furthermore, the dynamics of the concentration scaled to the respective minimum and maximum values were very similar in all catchments for single events with similar rain patterns (one example shown in Fig. 5) . Obviously, the characteristics of the three catchments causing different atrazine loads had little effect on the relative loss dynamics (Fig. 5). It seems that the dynamic aspects of the loss behavior were largely governed by the rainfall pattern, whereas the quantities of the herbicides lost were strongly influenced by catchment or field properties. This hypothesis is further supported by the observation that surface runoff and preferential flow into tile drains could hardly be distinguished based on their export dynamics monitored in Catchment 1 during a different year (Leu et al., 2004b).



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 5. (A) Cumulative rain amount of one event within the three catchments. (B) Relative discharge, (C) relative atrazine concentration, and (D) cumulative load caused by the event at the outlets of the catchments. The concentration and discharge values are scaled between the respective minimum and maximum levels of the event.

 
Nevertheless, for the periods between discharge events, we noticed a slight difference in the concentration dynamics among the three catchments. At the outlets of Catchments 2 and 3, the atrazine concentrations decreased to smaller levels compared with Catchment 1, where they remained rather high even for days after an event (Fig. 2, 6) . Because atrazine was mainly transported to the brooks by fast transport mechanisms, such a prolonged "tailing" was not expected. Seemingly, atrazine in Catchment 1 continued to leach from a short-lived pool in the saturated part of the soils after fast transport itself had stopped. The build-up of such a pool could be explained by bank storage. Squillace et al. (1996) observed stream water with high herbicide concentration moving into an adjacent alluvial aquifer (bank storage) during increased discharge. After the runoff event, this contaminated ground water discharged into the stream leading to a similar tailing effect in the stream.



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 6. Atrazine concentration exceedance curves at the outlets of Catchments 1 to 3 during a period of 60 d, beginning June 1 (start of the main application period).

 
Representativeness of the Catchments
In 1999, the periods previous to and during the application period were very wet. As a consequence, the maximum relative loss of atrazine of the entire 14-yr sampling period for the watershed of Lake Greifensee (1990 until 2003) was recorded (data not shown). Thus, the monitored losses from the three catchments, which cover about 3% of the area of the entire watershed of Lake Greifensee, are likely to be above average as well. These losses amounted to 2.1% of the applied amount. During the same period (1 June–31 July 1999), a total of 2.4% of the atrazine amount applied within the entire watershed was lost to the lake. This value includes contributions of several WWTPs that discharged 11% of the total atrazine input during this period (data not shown). The remaining 89% amounted to 2.1% of the atrazine applied. This relative input was identical to the relative losses from the 34 cornfields of the three study catchments. This match and the minor contributions from point sources (<11% of total losses on small and large scale) indicate that diffuse losses contributed a major part to the atrazine load to Lake Greifensee.

The representativeness of the three study catchments can also be assessed in a broader sense. The observed average atrazine losses are (slightly) higher than the mean value of 1.7% and the median of 0.47% compiled by Capel and Larson (2001) for catchments of very different sizes (100–100000 ha). For this comparison, it has to be considered that the study year was extremely wet. Hence, it can be expected for a normal meteorological year that losses from the study catchments are within the range observed in other areas. However, one has to take into account that reported losses may be due to diffuse losses as well as due to farmyard runoff. Actually, it has been observed that pesticide losses may be dominated by point sources. In the Zwester Ohm catchment (Germany), for example, Müller et al. (2002) found that point sources (WWTP and combined sewer overflows) contributed about 77% of the total input of 19 analyzed pesticides into surface water. However, the mean annual precipitation in that German catchment (585 mm) is less than half of the value in the Greifensee area (between 1160 and 1340 mm). Therefore, it comes as little surprise that the main reason for the different relative contributions from the two sources are the much lower diffuse losses observed in the dry German catchment. This can be illustrated by the example of isoproturon [3-(4-isopropylphenyl)-1,1-dimethylurea], which contributed >50% to the total pesticide pollution in the Zwester Ohm. Diffuse isoproturon losses from the Zwester Ohm catchment were 30 times smaller (0.07%) than the diffuse atrazine losses in the Greifensee catchment (2.1%.). The point source losses, on the other hand, differed very little (Greifensee catchment, ~0.3% of the atrazine amount applied; Zwester Ohm, 0.16% of the isoproturon amount applied). This observation shows that it is important to consider all major controlling factors including climate, soils, or land management when addressing the relative importance of point vs. diffuse sources across different regions.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The atrazine concentrations monitored at the outlets of the three catchments exhibited very strong fluctuations with concentration peaks exceeding 1 µg L–1 during periods of hours to a maximum of 2 d. Direct farmyard runoff caused the highest atrazine concentrations (up to 31 µg L–1) in the brooks. The losses from the fields resulted in somewhat lower concentrations but dominated the herbicide loads. The large concentration fluctuations are not only relevant for monitoring the herbicide losses but also for the assessment of the ecotoxicological effects of the herbicide contamination. Traditional tests simulate static exposure of aquatic organisms (e.g., algea) and may not be appropriate to mimic the real-world situations in small streams draining agricultural areas.

Despite the different relative losses observed at the outlets of all three catchments, the concentration dynamics were similar at all sites. It seems that the relative atrazine losses were strongly influenced by the characteristics of the cornfields whereas the travel times of these substances were hardly affected by the catchment properties. Therefore, we can infer little from the observed dynamics on the processes controlling the amounts of herbicides lost by fast transport mechanisms to surface waters. On the other hand, the knowledge of the dynamics in one catchment of regions with similar rainfall conditions may give a good indication for the dynamics to be expected in other catchments.

The herbicide losses from three different catchments varied substantially under very similar weather conditions. The observed differences confirm that within that type of landscape the variability of intrinsic field properties is a crucial factor controlling the quantitative herbicide losses to surface waters. This knowledge should be considered in catchment management approaches to reduce herbicide losses to surface water (e.g., avoiding applications to risk areas). Similar approaches have been proposed for other agrochemicals like P (e.g., Pionke et al., 2000; Walter et al., 2000). To delineate such risk areas, distributed models are needed that integrate the influence of all relevant field characteristics within small catchments. The data sets presented in this paper containing high resolution loss data as well as detailed spatial information on the herbicide application constitutes an interesting test case for such models.


   ACKNOWLEDGMENTS
 
The financial support of the Syngenta Crop Protection AG is gratefully acknowledged. We are indebted to M. Berg, M. Emmenegger, P. Fässler, M. Gehriger, A. Gerecke, G. Goudsmit, S. Heberle, A. Lück, J. Mühlemann, S. Oellers, M. Schärer, C. Stengel, and H. Wunderli (EAWAG and ETHZ) for the preparation and the analysis of the samples as well as for fieldwork. We thank M. Zahno, T. Ulaga, and M. Schenker for the GIS-based data analysis and C. Steiner for preparing the final version of Fig. 1. We are deeply grateful to A. Meerstetter and G. Popow (Kantonale Zentralstelle für Pflanzenschutz, Strickhof) for their contributions to the study design. We are thankful to all farmers of the catchment for their collaboration and to two anonymous reviewers for their helpful comments.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via ISI Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Leu, C.
Right arrow Articles by Stamm, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Leu, C.
Right arrow Articles by Stamm, C.
Agricola
Right arrow Articles by Leu, C.
Right arrow Articles by Stamm, C.
Related Collections
Right arrow Ecological Risk Assessment
Right arrow Organic Compounds
Right arrow Agricultural Pesticides
Right arrow Water Pollution

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal