Pesticides in stream water within an agricultural catchment in southern Sweden, 1990–1996

The Science of the Total Environment 216 Ž1998. 227]251

Pesticides in stream water within an agricultural catchment in southern Sweden, 1990]1996 J. Kreuger U Di¨ ision of Water Quality Management, Department of Soil Sciences, Swedish Uni¨ ersity of Agricultural Sciences, P.O. Box 7072, S-750 07 Uppsala, Sweden Received 10 February 1998; accepted 5 March 1998

Abstract Pesticide loss to stream water was studied in a small agricultural catchment in southern Sweden during the period 1990]1996. A total of 38 pesticides were detected in water samples, including 30 herbicides, four fungicides, three insecticides and one metabolite of one of the herbicides. Concentrations of pesticides in stream water were observed throughout the sampling periods. Peak concentrations occurred during the spraying seasons and following runoff events. Daily average concentrations sometimes varied by one order of magnitude from one day to another. Pesticides were also found in water samples as a result of incautious actions during handling and application procedures. Concentrations were lower at the outlet of the catchment area when the water had passed an open part of the stream, compared to concentrations detected in discharge from a culvert system upstream. This was largely a result of dilution from groundwater intrusion during low-flow periods. Sampling at different sites along the culvert demonstrated that the small village situated in the catchment did not contribute to pesticide findings in the culvert discharge. Wind drift had little influence on stream-water quality. Pesticide application for weed control in farmyards resulted in a substantial contribution to the pesticide load in stream water. Pesticides were persistent in the discharge throughout the winter and originated from both autumn and spring applications, as well as from farmyard application. Some autumn applied pesticides prevailed in stream flow during the following summer. Total amounts of pesticides lost in stream flow during May]September each year varied between 0.5 and 2.8 kg during the 7-year period, corresponding to ; 0.1% of the applied amount. Losses of single pesticides were generally less than 0.3% of the applied amount during individual years. Pesticides from agricultural applications in the catchment constituted, on average, 82% of the total transported amount lost during May]September each year, of which 2% was from autumn application the previous year. There was an overall correlation between amounts used in the catchment and occurrence in the water samples. The total pesticide load in water decreased markedly during the course of the investigation, in accordance with decreased amounts applied during spring and early summer. The results indicate that concentrations of some pesticides entering head-water streams in agricultural areas are close to, and during

U

Tel.: q46 18 672462; fax: q46 18 673430; e-mail: [email protected]

0048-9697r98r$19.00 Q 1998 Elsevier Science B.V. All rights reserved. PII S0048-9697Ž98.00155-7

228

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

certain time periods even above those levels demonstrated as having an impact on the aquatic flora and fauna. Q 1998 Elsevier Science B.V. Keywords: Herbicides; Fungicides; Insecticides; Surface water; Monitoring; Transport loss; Water quality; Watershed; Contamination

1. Introduction Monitoring programmes throughout Europe and North America have demonstrated a widespread presence of pesticides in streams and rivers ŽFrank et al., 1982; Baker and Richards, 1989; Clark et al., 1991; Larson et al., 1995; Lundbergh et al., 1995; Griffini et al., 1997.. Public concern is focused on possible negative impacts of pesticides on aquatic life or on human health. Proper estimation of any hazard that pesticides may pose to aquatic life is dependent upon knowledge of both exposure and toxicity. For adequate exposure assessment, as part of a risk evaluation, good quality data is needed on pesticide exposure patterns and characteristics ŽVan der Linden, 1994.. The ecological effects of pesticides on flora and fauna in surface waters are dependent on both peak concentrations and the duration of the exposure. Transport of pesticides from cultivated fields to surrounding surface waters generally occurs through runoff or drainage and is induced by rain or irrigation ŽLarson et al., 1997.. A number of field studies have investigated these mechanisms under a range of environmental conditions and for numerous pesticides. Measured losses were often in the range 0.01]0.5% and with maximum concentrations some times above 50 m grl Že.g. Traub-Eberhard et al., 1994; Brown et al., 1995.. However, also wind drift from nearby spraying applications, as well as incautious actions during the handling of pesticides, have been suggested as possible sources for pesticides entering stream waters ŽHarris et al., 1991.. Pesticide concentrations in streams and rivers leaving larger catchments are generally much lower than those in runoff water from field plots. Nevertheless, transported amounts in percentage of that applied are often independent of the size of the study area ŽBurgoa and Wauchope, 1995.. There

is a great need to increase our knowledge of transport pathways within a catchment influencing stream water quality. In most monitoring studies there has been a lack of site-specific data relating pesticide occurrence in the water with on-going activities in the catchment area. Only a few studies have been carried out on a catchment scale, with the aim of relating pesticide application timing and stream flow patterns to the occurrence of pesticides in the stream ŽMatthiessen et al., 1992; Jaynes et al., 1994; Laroche and Gallichand, 1995.. A pesticide monitoring study was initiated in spring 1990 to examine the loss of pesticides from an agricultural catchment in southern Sweden under normal management practices. Long-term studies within the catchment had demonstrated that losses of nutrients in drainage discharge and stream flow were in accordance with those measured from other catchments in southern Sweden. The catchment was thus selected as being typical for the region. Intensive use of pesticides in the area increased the possibility of generating information on the fate of a range of pesticides. The overall objective of this study was to quantify pesticide concentrations and mass transport in stream flow from a typical agricultural catchment in southern Sweden and to relate these concentrations to use patterns and transport pathways within the catchment. 2. Materials and methods 2.1. Catchment description The Vemmenhog ¨ catchment is located in the very south of Sweden Ž558 269 N; 138 279 E. on the south-western plain of Skane ˚ ŽScania province.. It forms the upper reach of the Vemmenhog ¨ Stream drainage basin which drains into the Baltic Sea.

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

The catchment, at an altitude of about 45 m above sea level, has undulating topography with glacial till-derived soils rich in chalk. The total thickness of the Quaternary deposits is 60]120 m ŽDaniel, 1992.. The total catchment area is 9 km2 Ž900 ha. consisting of 95% arable land, with sandy loam soils dominating. On average, the surface soil contains 2.5% organic matter Žwith a range of 1.9]2.8% between different sites. and the sandy loam consists of 55% Ž51]59%. sand, 29% Ž26]32%. silt and 16% Ž12]21%. clay. Sand, silt and clay contents are relatively constant in depth, but organic matter decreases to 0.1% at 1 m depth. There are a few small depressions within the catchment Ž; 5% of the total area. where soils are richer in organic matter and have a higher clay content. A qualitative analysis of clay minerals in the till from this region by X-ray diffraction showed that illite is the dominating fraction, but with a significant portion of smectite ŽKirchmann and Eriksson, 1993., giving the soils a slight swelling-shrinking capacity. The climate in the region is maritime due to the prevailing westerly winds and the vicinity of the Atlantic Ocean. The mean annual temperature is 7.28C, with mean summer and winter temperatures of 16 and y18C, respectively. The length of the growing season Žmean daily temperature ) 58C. is around 220 days. The average annual precipitation is 662 mm and falls mainly as rain. Soil temperature at 50 cm depth is 168C during the summer and drops to around 28C during the winter. Extensive drainage systems have been installed within the area and during the late 1950s open ditches in a larger part of the catchment were covered and replaced by a culvert system. About 40% of the field area is systematically drained with tiles spaced at 16 m and at a depth of ; 1 m. On the remaining area tile drains are installed in a non-regular manner following the natural drainage routes and connecting isolated depressions to the culvert system. The culvert collects tile drainage and also runoff water from surface runoff inlets Žvertical tiles. often used as inspection wells and located in the lowest-lying positions in the landscape along the tile drains in the field. Surface runoff inlets can also be found

229

along roads and in some farmyards. The culvert water is discharged into a small open stream that stretches 1.1 km to the outlet of the catchment. Just north of the outlet the stream has been widened to form a small, 0.2]0.4-m deep, pond. The pH-value of drainage and stream water is within the range 7.5]8.0. Average annual water flow volume within the catchment is 262 mm Žwhich amounts to 40% of the average annual precipitation. based on 20 years of flow measurement from a field site within the catchment. In general, the water flow rates are low for several months during the summer, with sudden peaks in the hydrograph in response to intense rainfall. The catchment is located in a typical cereal growing area, with an agricultural cropping practice normally comprising of a 4-year rotation with winter rape, winter wheat, sugar beet and spring barley on an average of 80% of the area. On the remaining area the following crops were grown: spring wheat, winter barley, oats, grass ley, peas, winter rye, rye wheat and spring rape. None of the crops were irrigated. There are 23 farms within the catchment area, cultivating 85% of the arable land, and another 12 farmers live at some distance from the catchment boundaries manage the remaining 15%. A small village, and also some scattered houses, are located within the catchment boundaries, and amount to a total of about 50 non-farming households. 2.2. Pesticide applications and usage Information on crops, fertilization, pesticide handling and usage on the field scale Ži.e. type of pesticides, dosage, time of spraying and field site location and size. was collected annually through personal interviews with the farmers. The nonfarming units were also interviewed during the first year of the investigation concerning pesticide usage. However, less than 10% of the non-farming households used pesticides, and these were mainly herbicides in small amounts, in their gardens. The total amount of pesticides applied within the catchment during each crop rotation was, on average, ; 1300 kg of active ingredient ŽAI., with 25% applied in the autumn Žend of August to

230

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

mid-October. and 75% during springrearly summer Žbeginning of April to early July.. On average, ; 60% of the total amount of pesticides applied each springrearly summer was applied during May. The pesticide usage during springrearly summer was dominated by herbicide applications Ž81% by weight., with the remainder made up of fungicides Ž15%., insecticides Ž3%. and growth regulators Ž1%.. Autumn applications were completely dominated by herbicides Ž) 99%.. During the 7-year period, 53 different AIs were applied on fields within the catchment, with an average annual use of 31 different AIs. Ten of these AIs accounted for 85]90% of the total weight applied. The application history for the most extensively used pesticides in the catchment is listed in Table 1. A complete record of pesticide applications in the catchment during 1990]1996 has been reported elsewhere ŽKreuger, 1996, 1997.. Of the total catchment area, on average, ; 80% was treated with pesticides during spring and early summer and ; 30% during the autumn, at an average rate of 1.5 kgrha and 1.3 kgrha, respectively. The application was highest during the first year and then dropped due to the introduction of the stereoisomers of dichlorprop and mecoprop Ždichlorprop-P and mecoprop-P. and

the sulfonylurea herbicide tribenuron-methyl, which are all used in low doses. This led to decreasing amounts of pesticides used in cereals during the investigation period. The most intensive usage of pesticides was in sugar beet which are grown on ; 20% of the area and receive almost half of the total amount of pesticides used during spring and early summer each year. The area treated with fungicides and insecticides varied greatly between years as a result of different weather conditions, with only 25% of the area treated during certain years and with up to more than 60% during other years. Most farmers living within the catchment applied the herbicide glyphosate for weed control outside cultivated areas, mostly in farmyards, but also along roads and around field edges, poles and surface water inlet wells. In addition, pesticides not registered for use in farmyards, e.g. terbuthylazine, simazine and cyanazine, as well as pesticides for which the approval of sale had expired, i.e. atrazine and dichlobenil, were also occasionally used. 2.3. Data collection and e¨ aluation 2.3.1. Sample collection Water sampling was started in spring 1990 and

Table 1 Most extensively used pesticides in the catchment 1990]1996 Pesticide

Type

Total use 1990]1996 Žkg.

Use Žmin]max. Žkgryear.

Area Žmin]max. Žharyear.

Average dose Žkgrha.

Springrearly summer Metamitron MCPA Chloridazon Fenpropimorph Dichlorprop-Pa Mecoprop-Pa Phenmedipham Propiconazole Ethofumesate Pirimicarb

Herbicide Herbicide Herbicide Fungicide Herbicide Herbicide Herbicide Fungicide Herbicide Insecticide

2024 849 744 678 509 506 333 229 177 157

147]470 42]234 10]287 41]150 13]188 19]181 23]68 14]52 10]35 3]57

78]189 71]307 12]141 224]495 37]255 30]269 78]201 221]495 53]162 18]463

2.10 0.57 1.70 0.27 0.46 0.76 0.35 0.09 0.23 0.12

Autumn Metazachlor Isoproturon Mecoprop Glyphosate

Herbicide Herbicide Herbicide Herbicide

819 672 216 182

25]225 16]234 0]81 2]49

18]214 16]246 0]126 4]53

1.16 0.95 0.80 0.95

a

Including the use of the non-stereoisomers dichlorprop and mecoprop, respectively, during the first years of investigation.

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

continued until autumn 1996, with the main sampling seasons during May]September each year. Water samples were initially collected from five sites within the catchment ŽFig. 1.: from a tiledrained agricultural field at site NA1 Ž32 ha., from the culvert at sites FA3 Ž482 ha. and SH5 Ž563 ha., at the outlet of the culvert at site UT10 Ž828 ha. and at the outlet of the catchment at site LU12 Ž902 ha.. During May]June 1990 and 1991, water samples were collected manually at all sites by periodically taking non-composite samples from the stream flow. Using programmable automatic samplers ŽISCO W models 2700, 2700R and 3700FR, ISCO W Inc., Nebraska, USA. time-paced sam-

231

pling was carried out at site LU12 during May]September 1990]1992 and at site UT10 during May 1992]June 1993, May]October 1994 and May]November 1995]1996. With the automatic samplers, time-paced samples were collected at daily or weekly intervals, each sample being a composite of sub-samples taken at 10-min or hourly intervals, respectively. All composite samples collected after mid-June 1990 were stored at q48C during the collection period in a sampler refrigerator. Samples were collected in glass bottles, prewashed with ethanol. To inhibit microbial degradation of the pesticides, dichloromethane was added to the bottles in advance, plus distilled

Fig. 1. Location of sampling sites and measurement devices in the catchment area.

232

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

water to prevent evaporation losses of the dichloromethane. After completion of the sampling programme, the bottles, capped with Teflon-lined screw caps, were delivered by mail to the laboratory within 48 h and extracted within 24 h of delivery. 2.3.2. Measurements Water outflow from the catchment area was continuously recorded by a water level gauge situated 200 m downstream from sampling location LU12. In April 1991, a Campbell datalogger was installed at the LU12 sampling location for continuous water flow recordings using a submerged probe. Culvert flow rates at UT10 were measured using a 90-degree V-notch weir and an ultrasonic sensor ŽISCO W model 3210 flow meter., starting in April 1993. Culvert flow rates before this date were calculated using a flow ratio between UT10 and LU12 based on measurements obtained after April 1993. Drainage discharge at the field site NA1 was quantified in a subsurface measuring station using a triangular weir and continuous recording by a water level gauge. Lack of flow data during short periods due to technical problems with any of the measurement devices was overcome by retrieving data from the closest possible measurement device, and making adjustments for differences in the scale factor. Rainfall was recorded on a daily basis at an official meteorological station ŽSkurup. located 6 km to the north-east of the catchment. Measurements made within the catchment during the summer months showed that precipitation amounts were similar to those at Skurup and amounted to 93]95% of the precipitation at Skurup. 2.3.3. Calculations Pesticide loads were calculated using measured concentrations in samples obtained by the automatic samplers and stream flow data as follows: Loads Q= c i = t i where Q is the mean stream flow Žlrs. during the collection period for each composite sample, c i is the observed concentration in the composite sam-

ple Ž m grl. and t i is the time represented by each sample Žs.. Pesticides not detected were assigned a value of zero for the load calculations. This means that loads and loss percentage estimates could be underestimated, since pesticides can be present in the water at a concentration just below the limit of determination. Since sample intervals varied during the season, mean concentrations were calculated weighting individual samples in relation to the time they represented. These time-weighted mean concentrations ŽTWMCs. were calculated as follows: TWMCs

Sc i t i St i

where c i is the observed concentration and t i is the time represented by each sample ŽRichards and Baker, 1993.. 2.3.4. Quality assurance The possible loss of unstable compounds during the collection and transport procedure was evaluated in two separate studies, using 14 pesticides representing different intrinsic properties. In the first study, spiked surface water samples were run through the automatic water sampler ŽISCO model 2700R. in the field, testing the stability during the collection procedure Ži.e. passage through the tubes of the sampling equipment. and during storage in the sampler ŽKreuger, 1992.. In the second study, pesticide stability was tested under different storage conditions, Žstorage time, storage temperature and following dichloromethane addition. ŽKreuger, 1994.. The general conclusion drawn from these studies was that the sampling and storage procedures used were adequate, with recoveries within the normal variance of the analytical procedures. 2.4. Analytical methods Pesticide residue analyses were conducted by the Department of Environmental Assessment, Organic Environmental Chemistry Unit, at the Swedish University of Agricultural Sciences. The 2.5-l unfiltered water samples were analyzed using two different procedures, the phenoxy acid

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

method and the multi-residue method. Special effort was made to include, in the analyses, as many as possible of the pesticides used within the catchment, with a maximum of 48 pesticides Ž33 herbicides, four fungicides and 11 insecticides . and two metabolites being screened for during the 7-year period. Twelve of the pesticides used in the catchment Žincluding glyphosate., representing 5% of the total volume applied, were not included in the analyses on any occasion due to analytical difficulties or to financial constraints when separate analytical methods were required. Some pesticides were included during the course of the investigation when adequate analytical methodology had been developed, e.g. chloridazon and ethofumesate. Samples were spiked with surrogate analytes to monitor the accuracy and precision of the analytical procedures. A limit of determination was calculated for every pesticide in each sample. All analytical results were compiled into the summary statistics, regardless of whether some of the analytical results fell below the stipulated limit of determination. These values are consequently not quantified with the normal precision but have been confirmed and constitute the best available estimates of the actual concentrations. It has been argued by other authors that biases can enter statistical summaries if low level data are censored in studies such as this ŽRichards and Baker, 1993.. Phenoxy acids and related compounds were hydrolyzed with alkali for 15 h at room temperature. After acidification, the acids were extracted with dichloromethane. Extractive alkylation with pentafluorobensylbromide and gas chromatogra˚ phy were conducted, as described by Akerblom et al. Ž1990.. Confirmation was by gas chromatography and mass spectrometry ŽGC-MS., with a limit of determination Žsignal-to-noise ratio s 3. in the range of 0.1]0.3 m grl. 2,4,5-TP was added as an internal standard from 1991 and concentrations were corrected according to extraction efficiency of 2,4,5-TP. Pesticides analyzed by this method were bentazone, bromoxynil, clopyralid, 2,4-D, dichlorprop, flamprop-M, fluroxypyr, ioxynil, MCPA and mecoprop. The recovery efficiency was within the range 90]102%, except for

233

clopyralid, fluroxypyr and ioxynil which had 60%, 75% and 150% recoveries, respectively, and with a standard deviation of 9]16% Žexcept ioxynil which had a standard deviation of 65%.. With the multi-residue method, semi-polar and non-polar pesticides were extracted with dichloromethane. Hydrophobic gel permeation clean-up and capillary column gas chromatography with selective detectors ŽGC-NP and GC-EC. were conducted according to the method de˚ scribed by Akerblom et al. Ž1990. and Andersson and Ohlin Ž1986.. During later years gel permeation was restricted to samples rich in humus or chlorophyll. Confirmation was by GC-MS, with a general limit of determination in the range of 0.1]0.5 m grl. Ethion was added as an internal standard from 1991 to confirm that the extraction efficiency was adequate, or the analytical procedure was repeated. The recovery efficiency of most substances was within the range 75]100% with a standard deviation of 13]28%; with the exception of fenpropimorph and metamitron that both showed low recovery, ; 50%, during the initial years of the investigation. None of the reported results with either of the two methods were corrected for recovery efficiency. Tribenuron-methyl, only analyzed during May] June in 1993 and in 1994, was not included in either of the two analytical methods described. Analysis of tribenuron-methyl was performed using liquid-liquid extraction after pH-adjustment, followed by HPLC-MS. 3. Results and discussion 3.1. Rainfall and waterflow Annual precipitation ranged from 537 mm Ž1996. to 837 mm Ž1994. during the investigation period, with the largest variation in monthly precipitation between years during the summer months June and July ŽFig. 2.. As a whole, warmer and drier weather than normal was predominant during sampling periods. Flow volumes showed great fluctuations between years ŽTable 2.. During the cold and rainy spring and early summer of 1991 flow was quite high, whereas during the warm and dry spring and summer of 1992 it was

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

234

extremely low. Flow was closer to the average in 1990, but with thunderstorms in June and July giving sudden high flow events. In the beginning of 1993 there was very little precipitation for several months, resulting in a low-flow situation. However, immediately after the sampling was discontinued on the last day of June, there was a change due to cold and rainy weather resulting in substantial drainage discharge later that year. The flow situation during 1994 was close to normal apart from a high-flow event in May and also in

September, which received more than double the normal amount of precipitation. Both in 1995 and in 1996 there were long periods of low-flow situations, and total flow volume during 1996 was the smallest since the onset of the investigation. 3.2. Pesticide findings 3.2.1. Concentrations Pesticides were detected in almost every stream-water sample collected at sites LU12 and

Table 2 Monthly flow totals during 1990]1992 at site LU12 and 1992]1996 at site UT10, with sampling periods marked in bold type Month

Monthly flow Žmmrm. 1990rLU

January February March April May June July August September October November December Total a

29.8 21.0 32.9 12.3 10.8 14.3 7.5 5.0 14.4a 14.1 23.7 23.0 208.8

1991rLU

1992rLU

1992rUT

1993rUT

1994rUT

1995rUT

1996rUT

51.8 24.9 21.7 20.4 34.7 12.7 8.3 4.1 3.7 10.8 22.0 46.0

39.2 32.2 31.8 26.3 5.9 3.0 2.6 2.6 2.6 13.1 47.8 38.1

30.6 23.2 21.3 8.7 2.8 0.6 0.2 0.2 0.2 7.4 38.2 28.9

48.6 28.4 11.5 5.4 1.4 0.3 1.6 6.3 38.2 43.5 30.6 32.0

20.2 3.5 37.2 13.6 8.4 3.5 0.5 0.4 16.5 4.3 21.0 47.8

49.7 52.4 18.7 16.3 7.6 7.9 1.0 0.3 0.5 0.4 1.4 1.5

4.3 7.7 17.3 9.2 8.7 2.3 0.6 0.1 0.2 0.2 2.6 11.5

261.1

245.1

162.3

247.8

176.8

157.7

64.8

Sampling was interrupted in mid-September, actual flow volume during sampling period was 3.4 mm.

Fig. 2. Monthly precipitation totals during 1990]1996, with the 30-year monthly average precipitation indicated as a column.

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

UT10. The overall frequency and concentration levels at which pesticides were identified during the summer period, May]September, at sites LU12 1990]1992 and UT10 1992]1996 and also at site UT10 during the winter period, October]April, 1992]1996 are statistically summarized in Tables 3]5. A complete record of all findings has been reported earlier in two techni-

235

cal reports ŽKreuger, 1996, 1997.. Altogether, 38 pesticides, distributed among 30 herbicides, four fungicides, three insecticides and one metabolite, were identified. These varied from single events of benazolin-ethylester, dichlobenil, phenmedipham and triadimenol, to overall findings of more than 50% detection frequency for the herbicides atrazine, bentazone, dichlorprop, ethofumesate,

Table 3 Summary statistics of detections and concentration levels in water samples collected at site LU12 during May]September 1990]1992 Pesticide

Herbicides Atrazine Benazolinethylester Bentazone Bromoxynil Chloridazon Clopyralid Cyanazine 2,4-D Dicamba Dichlobenil Dichlorprop Diuron Ethofumesate Flamprop-M Fluroxypyr Ioxynil Isoproturon Lenacil Linuron MCPA Mecoprop Metamitron Metazachlorf Methabenzthiazuron Pendimethalin Phenmedipham Simazine Terbuthylazine

Use rank in areaa

No. of samples

Detections Ž%.

LODb Ž m grl.

f.y.e

115 66

41% 0%

0.1 0.2

15 18 3

10

115 38 30 115 115 115 85 115 115 115 64 65 65 65 115 85 115 115 115 115 115 115

73% 13% 23% 0% 11% 12% 0% 0% 63% 0% 45% 6% 3% 17% 10% 0% 3% 61% 61% 28% 51% 7%

0.1 0.2 0.7 0.3 0.1 0.1 0.1 0.1 0.1 0.5 0.2 0.1 2.0 0.2 1.0 0.5 0.5 0.1 0.1 1.0 0.1 0.3

20 7 f.y. f.y.

115 84 115 115

0% 0% 2% 67%

0.2 2.0 0.1 0.1

Fungicides Fenpropimorph 5 Prochloraz 13 Propiconazole 8 Triadimenol

115 62 115 115

4% 0% 40% 0%

0.3 0.2 0.2 0.5

19

f.y. 4 12 16 11 17

2 6 1

Concentration Ž m grl. Max TWMC c

9.3

Max weekly TWMC

Max monthly TWMC d

Percentile of weekly TWMC 90th

75th

50th

25th

0.9

0.5

0.3

0.1

0.02

0

10 1 10

3.8 0.9 10

1.3 0.3 2.1

0.4 0.05 1.0

0.13 0 0.5

0.04 0 0

0 0 0

2.8 10

1.1 10

2.3

0 0.01

0 0

0 0

0 0

20

20

4.6

0.7

0.3

0.05

0

0.9

0.3 0.7

0.3 0 0 0.08 0.09

0.1 0 0 0 0

0 0 0 0 0

0 0 0 0 0

2.0 0.05 7.0 1.0 2.0

2.0 0.05 7.0 0.9 1.8

1.8 40 7.0 45 5.1 0.6

0.6 18 6.0 24 3.6 0.2

4.7 2.2 7.1 1.3 0.1

0 1.1 0.4 1.6 1.0 0.2

0 0.4 0.1 0.2 0.2 0

0 0.06 0.04 0 0 0

0 0 0 0 0 0

0.3 2.0

2.0

0.7

0.4

0.3

0.1

0

1.0

1.0

0

0

0

0

2.8

1.3

0.6

0.2

0

0

1.0

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

236 Table 3 Ž Continued. Pesticide

Insecticides Alphacypermethrin Cyfluthrin Cypermethrin Deltamethrin Esfenvalerate Fenitrothion Fenvalerate Permethrin Pirimicarb Sum of pesticides

Use rank in areaa

9

No. of samples

Detections Ž%.

LODb Ž m grl.

115

0%

0.2

115 115 115 66 115 115 115 115

0% 0% 0% 0% 0% 0% 0% 36%

0.5 0.2 0.1 0.2 0.1 0.2 0.5 0.1

115

97%

Concentration Ž m grl. Max TWMC c

2.0 124

Max weekly TWMC

1.0 64

Max monthly TWMC d

0.3 19

Percentile of weekly TWMC 90th

75th

50th

25th

0.2

0.08

0

0

8.8

3.9

1.1

0.4

a

Rank order of spring applied amounts in the field during 1990]1992. Median limit of determination ŽLOD.. c Time-weighted mean concentration ŽTWMC.. d Only calculated if detected in water samples for a whole month. e Farmyard application Žf.y... f Autumn application of metazachlor often starts in mid-August. b

MCPA, mecoprop, metazachlor and terbuthyazine. Many of the pesticides were found in water over several months, but with maximum concentrations occurring during or shortly after the application period. The results indicate an annual cycle of pesticides entering the stream water during periods of application, followed by a series of flushing events during storm-flow periods. These findings support the concept of a rapid transport mechanism in areas with subsurface drainage networks which increase drainage either by facilitating the removal of infiltrating water or collecting runoff water and removing it through surface runoff inlets Žvertical tiles. along the tile drains in the field ŽSchottler et al., 1994.. The magnitude of concentration peaks varied to a great extent, most often due to changes in stream flow, with daily average concentration sometimes varying by one order of magnitude from one day to another. However, taking the results as a whole, there was no direct correlation between pesticide concentration and the magnitude of stream flow ŽFig. 3.. Laroche and Gallichand Ž1995. made the same observation in a Canadian catchment study.

The highest, time-weighted, mean concentrations ŽTWMC. measured during a single week were for the herbicides metamitron Ž24 m grl. at LU12 and metazachlor Ž200 m grl. at UT10. Concentration ranges of major pesticides which occurred in stream flow at site UT10 during the summer months are graphically presented in Fig. 4. Median concentration values for most of these pesticides were 0.1]0.2 m grl, apart from terbuthylazine which had a median concentration of one order of magnitude higher and a total pesticide median concentration at two orders of magnitude higher. Looking at the results as a whole, there was a good correlation during the individual years between amounts used and the concentrations found in the water samples ŽKreuger and Tornqvist, 1998.. Pesticides that were not de¨ tected, or detected only on single occasions, were mostly applied in small amounts, or else a higher limit of determination could explain such non-detection. Because of the higher limit of determination for some of the investigated pesticides, it is likely that these had undetected concentrations in the same order as those encountered for other

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

compounds. For example, phenmedipham, analyzed at a limit of determination 10 times above that of most other compounds, was only detected on a single occasion despite regular use in the area. Also findings of metamitron and chloridazon, as well as isoproturon in 1990]1991, might

237

have been underestimated due to elevated limits of determination. Generally, this problem was likely less pronounced during years of large quantities applied in the area. As a whole, concentrations were lower at the outlet of the catchment ŽLU12., where the water

Table 4 Summary statistics of detections and concentration levels in water samples collected at site UT10 during May]September 1992]1996 Pesticide

Herbicides Atrazine Atrazine-desethylf BAMg Benazolin-ethylester Bentazone Chloridazon Clopyralid Cyanazine 2,4-D Dichlobenil Dichlorprop Diflufenicanh Diuron Ethofumesate Flamprop-M Fluroxypyr Hexazinon Ioxynil Isoproturon MCPA Mecoprop Metamitron Metazachlori Methabenzthiazuron Pendimethalin Phenmedipham Propyzamide Prosulfocarbh Simazine Terbuthylazine Tribenuron-methylh Fungicides Fenpropimorph Prochloraz Propiconazole Triadimenol

Use rank in areaa

f.y.e

2 f.y. f.y. 7 f.y. 9 12 f.y. 15 10 4 5 1 16 6 f.y. f.y. f.y.

3 19 8

No. of samples

Detections Ž%.

LODb Ž m grl.

95 51 51 95 97 95 97 95 97 95 97 7 95 95 67 87 48 33 95 97 97 95 95 95 95 66 95 7 95 95 30

72% 6% 0% 2% 76% 48% 13% 26% 21% 1% 65% 100% 6% 85% 36% 49% 21% 24% 61% 64% 78% 58% 83% 29% 0% 2% 5% 0% 45% 96% 87%

0.1 0.2 0.1 0.2 0.1 0.5 0.3 0.2 0.1 0.1 0.1 0.1 0.5 0.1 0.1 0.3 0.1 0.2 0.2 0.1 0.1 0.5 0.2 0.2 0.2 1.0 0.1 0.2 0.1 0.1 0.01

95 95 95 95

61% 9% 65% 1%

0.1 0.5 0.2 0.3

Concentration Ž m grl. Max TWMC c

Max weekly TWMC

Max monthly TWMC d

3 0.1

3 0.1

0.4 5 20 10 10 10 0.2 25 0.1 0.6 6 2 6 1 3 4 60 16 60 200 7

0.4 5 20 6 10 10

3.9 6.0 2.1 6.0

25

9.9

0.6 4 1 5 1 3 3 39 16 60 200 4

0.1 2.3 0.4 1.9

2 3

2 3

15 20 0.4

10 20

8 2 20 3

6 1 12

Percentile of weekly TWMC 50th

25th

0.3 0

0.1 0

0 0

0 0.9 2.0 0.02 0.6 0.1 0 2.0

0 0.3 0.6 0 0.08 0 0 0.4

0 0.1 0 0 0 0 0 0.1

0 0.01 0 0 0 0 0 0

0 1.0 0.1 0.9 0.1 0.4 0.7 4.3 3.3 9.1 2.0 0.5

0 0.8 0.03 0.3 0.05 0 0.3 0.9 0.9 3.7 0.5 0

0 0.2 0 0 0 0 0.1 0.1 0.2 0.3 0.2 0

0 0.09 0 0 0 0 0 0 0.07 0 0.1 0

0 0

0 0

0 0

0 0

5.4 11.7

2.0 6.0

0.2 3.0

0 1.0

0 0.3

2.7

0.9 0 2.0 0

0.4 0 0.9 0

0.1 0 0.1 0

0 0 0 0

1.5

1.0 9.5 4.2 15.6 48.5 1.4

5.2

90th

75th

0.7 0

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

238 Table 4 Ž Continued. Pesticide

Insecticides Cyfluthrin Cypermethrin Deltamethrin Dimethoate Esfenvalerate Fenitrothion Lambda-cyhalotrin Pirimicarb Sum of pesticides

Use rank in areaa

20

11

No. of samples

Detections Ž%.

95 95 95 42 95 95 51 95

7% 0% 0% 5% 0% 0% 0% 64%

100

100%

LODb Ž m grl.

0.2 0.2 0.1 0.1 0.1 0.1 0.1 0.1

Concentration Ž m grl. Max TWMC c

Max weekly TWMC c

Max monthly TWMC d

Percentile of weekly TWMC 90th

75th

50th

25th

5

5

0

0

0

0

30

30

0 0

0 0

0 0

0 0

10

7

2.0

0.5

0.07

0

213

213

61.0

24.8

9.6

3.6

3.9 63

a

Rank order of spring applied amounts in the field during 1992]1996. Median limit of determination ŽLOD.. c Time-weighted mean concentration ŽTWMC.. d Only calculated if detected in water samples for a whole month. e Farmyard application Žf.y... f Desethyl-atrazine, a metabolite to atrazine. g BAM, i.e. 2,6-dichlorobenzamide, a metabolite to dichlobenil. h Only analyzed during a short period of time Žtribenuron-methyl: May]June 1993 and 1994; others: starting in July 1996.. i Autumn application of metazachlor often starts in mid-August. b

had passed through the open part of the stream for 1.1 km, compared to concentrations detected in culvert discharge ŽUT10.. Water flow measurements from these two sites showed that there was a substantial contribution to stream flow from groundwater intrusion, especially during low flow conditions. Pesticide, as well as nutrient, concentrations were thus attenuated by around one order of magnitude between the two sites during the summer months. The difference in concentrations between these two sites was continuously investigated during the warm and dry summer of 1992. In mid-July, after 2 months with almost no rain, 27 mm of rain fell during a 3-day period. Some very high concentrations were found in culvert discharge during this period, whereas concentrations were considerably lower in corresponding samples collected at the outlet of the catchment. Apart from dilution, this could also be explained by the tendency for certain pesticides to distribute to sediment. Some of these pesticides Žnotably cyfluthrin, methabenzthiazuron and

prochloraz. were only detected in culvert discharge and not in the stream at the outlet of the catchment. The application period had stopped 1 month earlier for most of the pesticides, but, possibly, the very dry weather lowered degradation rates of pesticide residues in the topsoil during this period. A dried-out topsoil may also have allowed water flow in cracks, thus moving some of the pesticides rapidly through the topsoil to the tile drains. In a study by Bergstrom ¨ and Jarvis Ž1993. they found higher concentrations of dichlorprop in leachate from a clay soil under a low irrigation regime than under an intense irrigation regime, which they attributed to macropore flow in the soil treated with a small water input before leaching started. Pesticide analysis in this study was carried out on unfiltered water samples. The relative importance of pesticide transport dissolved in water or adsorbed onto suspended solids has been investigated by several authors ŽPereira and Rostad, 1990; Clark et al., 1991; Gomme et al., 1991;

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Brown et al., 1995. on a range of pesticides Žatrazine, chlortoluron, cyanazine, fonofos, isoproturon, mecoprop, metolachlor, propyzamide, simazine, tri-allate and trifluralin .. These results suggest that the greater part of the pesticide load is carried dissolved in water. Only for the low water soluble pesticide trifluralin, movement on sediment was shown to be the dominant mode of transport. This agrees with the observations of

239

Wauchope Ž1978., i.e. that sediment associated transport is the primary process only for pesticides with low water solubilities. Some of the concentrations detected, especially in culvert discharge, were much higher than those encountered in previous Swedish monitoring studies. However, similar concentrations of pesticides in runoff water have been found in a number of field studies elsewhere ŽBurgoa and Wauchope,

Table 5 Summary statistics of detections and concentration levels in water samples collected at site UT10 during October]April 1992]1996 Pesticide

Herbicides Atrazine Atrazine-desethylf BAMg Benazolin-ethylester Bentazone Chloridazon Clopyralid Cyanazine 2,4-D Dichlobenil Dichlorprop Diflufenicanh Diuron Ethofumesate Flamprop-M Fluroxypyr Hexazinon Ioxynil Isoproturon MCPA Mecoprop Metamitron Metazachlori Methabenzthiazuron Pendimethalin Phenmedipham Propyzamide Prosulfocarbh Simazine Terbuthylazine Fungicides Fenpropimorph Prochloraz Propiconazole Triadimenol

Use rank in areaa

f.y.e

f.y. 8 5 f.y.

f.y. 11 2 10 4 1 12

f.y. 7 f.y. f.y.

No. of samples

Detections Ž%.

LODb Ž m grl.

41 13 13 41 41 41 41 41 41 41 41 6 41 41 34 36 16 28 41 41 41 41 41 41 41 28 41 6 41 41

44% 15% 0% 0% 44% 7% 2% 0% 0% 0% 29% 100% 5% 24% 15% 17% 0% 7% 51% 10% 56% 5% 71% 22% 0% 0% 0% 67% 20% 98%

0.1 0.2 0.1 0.1 0.1 0.7 0.3 0.1 0.1 0.1 0.1 0.1 0.3 0.1 0.1 0.2 0.2 0.2 0.2 0.1 0.1 0.5 0.1 0.3 0.2 1.0 0.1 0.2 0.1 0.1

41 41 41 41

34% 0% 32% 0%

0.1 0.3 0.2 0.2

Concentration Ž m grl. Max weekly TWMC c

Max monthly TWMC d

Percentile of weekly TWMC 90th

75th

50th

25th

1 0.1

0.59

0.5 0.08

0.1 0

0 0

0 0

0.5 15 0.5

0.35

0.3 0 0

0.2 0 0

0 0 0

0 0 0

0.8 0.3 0.2 0.3 0.1 1.8

0.20

0.06

0.03

0

0

0.15 0.05 1.00

0 0.2 0.03 0.4

0 0 0 0

0 0 0 0

0 0 0 0

0 0.8 0 0.1 0 0.7 0

0 0.05 0 0.04 0 0.1 0

0 0 0 0 0 0 0

0.2 10 0.5 5 0.4 5 30

2.48 10.65

0 4 0 1 0 1 1

0.8 2 100

1.03 47.84

0.2 5

0 1

0 0.5

0 0.2

2

0.53

0.2

0.1

0

0

0.6

0.60

0.4

0.1

0

0

4.87 1.80

240

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Table 5 Ž Continued. Pesticide

Use rank in areaa

No. of samples

Detections Ž%.

Insecticides Cyfluthrin Cypermethrin Deltamethrin Dimethoate Esfenvalerate Fenitrothion Lambda-cyhalotrin Pirimicarb

41 41 41 13 41 41 13 41

0% 0% 0% 0% 0% 0% 0% 24%

Sum of pesticides

41

98%

LODb Ž m grl.

0.2 0.2 0.1 0.1 0.1 0.1 0.1 0.1

Concentration Ž m grl. Max weekly TWMC c

Max monthly TWMC d

Percentile of weekly TWMC 90th

0.4

0.22

0.2

145.7

65.84

18.26

75th

50th

25th

0

0

0

7.86

1.72

0.5

a

Rank order of autumn applied amounts in the field during 1992]1996. Median limit of determination ŽLOD.. c Time-weighted mean concentration ŽTWMC.. d Only calculated if detected in water samples for a whole month. e Farmyard application Žf.y... f Desethyl-atrazine, a metabolite to atrazine. g BAM, i.e. 2,6-dichlorobenzamide, a metabolite to dichlobenil. h Only analyzed autumn 1996. i Autumn application of metazachlor often starts in mid-August. b

1995.. Leonard Ž1990. concluded in a review article that runoff losses at the edge of the field may reach several percent of the amount applied and concentrations may reach several mgrl if runoff occurs soon after application. These runoff concentrations are, however, rapidly attenuated in the transport system by dilution, deposition and trapping of sediments along the flow path. Another factor to consider in catchment-scale monitoring is the time of application. Not all of the catchment is treated on the same day. On the other hand, the probability of one or several fields being newly treated with pesticides when rainfall starts increases in a catchment with many fields, managed by many different farmers. In larger catchments, rainfall and runoff are usually distributed in time and space such that an additional attenuation of pesticide loads in runoff at the outlet of the catchment would be expected ŽLeonard, 1990.. The non-composite samples taken at different sites along the culvert at weekly or bi-weekly intervals demonstrated that pesticides were found in subsurface flow at all sites. However, only

small amounts of pesticides were detected at the field site NA1. The sampling procedure being non-composite and not flow-event related was a drawback when monitoring pesticide transport from this small field site. Since pesticide fluxes in response to rainfall, at least from smaller fields, are most often of very short duration, non-composite sampling on a few occasions, would most likely tend to underestimate the pesticide transport Žunless sampling occasions were directed exclusively to runoff situations .. Pesticide transport studies at other field sites have demonstrated that the largest masses of pesticides are most often lost during periods of very short duration at the early part of a flow event ŽKladivko et al., 1991; Brown et al., 1995.. A small village is situated between sites SH5 and UT10 and nutrient analyses showed a marked increase in phosphorus concentrations between these two sites, especially during the summer months, as a result of discharge from households into the culvert. However, pesticide concentrations found in the culvert were of the same order of magnitude before ŽSH5. and after ŽUT10. the

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Fig. 3. Concentration Žweekly TWMC. of MCPA, ethofumesate and metazachlor, along with total pesticide concentration, in relation to stream flow volume in samples collected during 1994]1996 at site UT10. For single pesticides points are divided to show when the sample was collected in relation to application period, i.e. during application period, less than 1.5 months after last application and more than 1.5 months after last application.

village, indicating that pesticides from household use were not contributing to the findings in the stream, except atrazine which showed increased concentrations between these two sites. This was due to its application at two farmyards for weed

241

control, both of which were situated downstream from site SH5. Low concentrations of atrazine detected in water at sites FA3 and SH5 most likely originated from past uses. Buhler et al. Ž1993. found atrazine in tile drainage for up to 3 years after use of the compound was discontinued on the field, demonstrating that it may take several years before the impact of altered practices can be evaluated. Pesticides applied in the autumn were still detected in water samples collected the following spring and summer. For example, metazachlor was occasionally present at 0.05]0.5 m grl in water samples collected during the summer following an autumn application. Also isoproturon in 1996, prevailed in water samples at 0.07]0.4 m grl through the summer period, despite no spring application in the area. The persistence of isoproturon in soil water for up to a year after last application has been demonstrated under controlled conditions in other field studies ŽJohnson et al., 1996.. When application of metazachlor and isoproturon started in the autumn, concentrations increased markedly in the water samples. Water sampling during the winter season 1992r93, showed that both metazachlor and isoproturon were detected in winter discharge during extended periods. Also bentazone and mecoprop were occasionally detected at low concentrations in water samples after the winter period, prior to being applied in spring. Not all pesticides entering the stream had an agricultural use in the catchment. Findings of atrazine, hexazinon, propyzamide, simazine and terbuthylazine Žand also bentazone and cyanazine during certain periods. derived from their application on non-agricultural land, mainly for weed control in farmyards. One extraordinary example was the application of terbuthylazine at one farmyard for weed control in September 1992. Terbuthylazine was subsequently detected in culvert discharge with a maximum concentration of 100 m grl at the onset of increased discharge in midOctober, and was thereafter present throughout the winter season. This resulted in a total load of more than 6 kilos of terbuthylazine during a 7-month period. Atrazine, only applied in farmyards in this area, was detected in stream flow

242

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Fig. 4. Concentration ranges Žweekly TWMC. of major pesticides at sampling site UT10 during May]September 1992]1996.

throughout the entire investigation period. It is notable that atrazine was withdrawn from the Swedish market in 1989, but old stocks have been used regularly, although in decreasing quantities, during the entire investigation period. High concentrations of a wide range of compounds were detected in water collected in surface runoff inlets in two farmyards on three occasions in May]June 1991 and on one occasion in May 1992. Some very high concentrations were encountered, especially in water from a larger farm. On the last sampling occasion in 1991, there was a runoff event, due to heavy rainfall starting about 8 h before the sample was taken, and with a total of ; 40 mm of rainfall since the previous sample. Nevertheless, most of the previously detected compounds were still present at elevated concentrations in runoff water from the farmyard, demonstrating the possibility of extended duration of pesticide losses also from point sources within a catchment. Despite these elevated concentrations attributed to point sources, there was no obvious influence on the composi-

tion of pesticides in the culvert or stream water during the subsequent runoff events in 1991. There was no flux in the stream containing a similar composition of pesticides as detected in water from either well. In spring 1995, bromide was applied as a non-reactive tracer in the yard of the larger farm in an attempt to evaluate the possible impact on stream water quality of pesticide spills in farmyards. However, no bromide was detected in culvert discharge during the 8-week period following the application, despite some heavy rainfall during this time. Bromide is extremely mobile and might have leached into the soil profile rather than mixing into the runoff water. In a study by Hubbard et al. Ž1989. bromide was not affected by rainfall intensity, instead the bromide was essentially leached away from the runoff zone. The connection of elevated pesticide concentrations with farmyards is cause for concern, since the lack of a microbiologically active, soil surface layer at these sites increases the potential risk of rapid pesticide leaching to groundwater aquifers.

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Occasionally, elevated pesticide concentrations were found in the stream without association to any rainfall event. One such occasion was on 10 May 1990, when elevated concentrations were found both in culvert water Žtotal pesticide concentration of 346 m grl. and at the outlet of the catchment. There had been no rainfall during the preceding 10 days. Also, in May]June 1992 there were elevated concentrations in the culvert discharge and stream flow during a long period without any rainfall in the area. These incidents were most likely the result of spillage occurring in the catchment while filling or cleaning spraying equipment. 3.2.2. Transport and loss In the following presentation, the term transport will refer to the amount of pesticides leaving the catchment, whereas the term loss will refer to transported amounts as a percentage of applied amounts. Total amount of pesticides transported in stream flow during May]September varied between 0.5 and 2.8 kg during the 7-year period, with largest amounts normally lost during the application season May]June ŽFig. 5.. Although concentrations detected in culvert discharge were generally higher than those encountered in stream flow at the outlet of the catchment, calculated losses during 1992 for most compounds were comparable between these two sites. September 1994 was extremely wet with 150 mm of total rainfall. Following the autumn application, ; 1 kg of metazachlor was lost in water during this month, corresponding to 0.44% of the applied amount. Transported amounts during the summer months of the 10 most commonly used pesticides in the catchment during the 7-year period Žsee Table 1. constituted on average 63% of total transported amounts, with a yearly variation ranging from 32]83%. The total pesticide load in water decreased markedly during the investigation period ŽFig. 5.. In 1990, during May]September, a total of 2.5 kg of pesticide residues were transported in water leaving the catchment. The corresponding figure in 1996 was 0.5 kg. Pesticide load originating from non-agricultural application was, during most summers, in the order of 0.1 kg, with the excep-

243

tion of the summer of 1994, when ; 0.9 kg originating from non-agricultural application was lost. Information on spring applied amounts of those pesticides occurring in the water is included in Fig. 5. Comparatively more pesticides were lost during the first 2 years Ž; 0.2% of used amount., whereas the relation between transported and applied amounts was constant during the following years Ž0.06]0.1%.. Lost amounts have, however, decreased in absolute figures, in accordance with decreased amounts applied during spring and early summer. Losses for single pesticides were generally less than 0.3% of the applied amount during individual years and as a whole the average loss during May]September was ; 0.1% of the applied amount ŽFig. 6.. There was no apparent difference in loss between pesticides applied at different rates, which is in accordance with other studies ŽWauchope et al., 1990.. Occasional extreme losses of single pesticides were registered during the investigation and were attributed to pointsource contributions or non-agricultural applications. Pesticides with an agricultural application in the catchment constituted, on average, 82% of the total transported amount lost during May]September each year, of which 2% was from autumn application the previous year. The remaining 18% was made up of pesticides originating from non-agricultural uses. Total amounts of pesticides lost in stream flow during the winter period of 1992r1993 ŽOctober]April. was 8.9 kg, of which terbuthylazine constituted 6.2 kg. The remaining 2.7 kg was of the same order as maximum losses during the summer period. Although concentration levels were low during most of the winter period, transported amounts were substantial ŽFig. 7.. Losses of the phenoxy acids mecoprop and dichlorprop during the winter period after autumn application were in the same range as those estimated during the summer months Ž0.1]0.2%.. Isoproturon, on the other hand, showed a 100-fold increase in transported loss Žto 1.96%. after autumn application compared to losses after spring application Ž0.02%., even though it was only applied in spray mixtures with dichlorprop and MCPA to cereals in autumn 1992. Isoproturon is

244

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Fig. 5. Total amount of pesticides transported during May]September 1990]1996. In the top figure ŽA. the columns are divided to show the different time periods, May]June and July]September, and also include transport of metazachlor occurring in stream flow after its application in August. In the bottom figure ŽB. the columns are divided to show the origin of the pesticides transported, i.e. from agricultural spring application, from agricultural autumn application the previous year and from non-agricultural application. Information on spring applied amounts of those pesticides occurring in the water is included in ŽB..

typically more strongly sorbed and has a slower degradation rate than mecoprop. Harris et al. Ž1994. found isoproturon Žbut no mecoprop. in runoff water up to 3 months after autumn application, with total loss amounting to 0.5% of that applied. Most of this isoproturon loss to drain

flow was considered to be the result of bypass flow through cracks before the soil was wetted up. In another study, Harris et al. Ž1995. reported isoproturon losses from field plots in the range of 1.7 to 3.3% after autumn application to a heavy clay soil. Metazachlor, applied in AugustrSep-

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

tember 1992, was detected in culvert discharge throughout the winter, and amounted to 0.32% of the application. Some of the pesticides that were detected in water during the winter period 1992r93 were not applied during autumn, thus enabling transport calculations for some of the pesticides reflecting a 12-month period. The warm and dry summer of 1992 possibly reduced degradation and leaching of the pesticides, as many of the spring applied pesticides were detected in discharge when a major leaching period started in mid-October. The transported loss of pesticides during May]September 1992 was, on average, only 0.06% of the applied amount, which was considerably less than the loss measured during the 2 previous, wetter, years. However, when taking transported amounts measured during the winter period later that year into account, average loss increased to 0.15% of applied amounts, thus demonstrating that for several of the pesticides there was a delay in transported amounts reaching surface water. In a field study by Gaynor et al. Ž1992. a major part of atrazine losses was postponed until after harvest during drier years as opposed to normal years

245

when most losses occurred during the growing season. The most noticeable increase in transported loss during the winter period occurred for the low water soluble and persistent fungicide fenpropimorph. Fenpropimorph was detected in winter discharge, with a maximum loss during a 3-week period with intense rainfall and low temperatures during JanuaryrFebruary, about 7 months after the last application. Total losses of fenpropimorph during the 12-month period was equivalent to 0.3% of that applied. Brown et al. Ž1995. observed extended losses over the winter season for trifluralin, another low water soluble and relatively persistent pesticide. Their results showed that movement associated with sediment was the dominant mode of transport for trifluralin. Increased transport of this compound during late winter was attributed to increased transport of sediment associated with breakdown of soil aggregates by freezing and thawing cycles over the winter period. Burgoa and Wauchope Ž1995. summarized the literature on pesticide losses in runoff waters from agricultural fields and concluded that for

Fig. 6. Total seasonal loss in stream water at site LU12 Ž1990]1991. and site UT10 Ž1992]1996. for the most heavily used pesticides during May]September ŽMay]June in 1993.. Žn.a., not analyzed..

246

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Fig. 7. Total monthly pesticide concentration Žtop. and transported amount Žbottom. in stream water samples collected at site LU12 Ž1990]1991. and site UT10 Ž1992]1996..

the majority of pesticides total losses are 0.5% or less of the amounts applied. Most loss rates measured in the Vemmenhog ¨ catchment were in agreement with these conclusions. 4. Ecotoxicological significance Peak concentrations of pesticides during the summer months could be a threat to the flora and

fauna living in small headwater streams such as the Vemmenhog ¨ catchment. This might lead to less diverse communities if the most sensitive species are affected, but, on the other hand, less sensitive species could possibly increase in number. In a study on the toxic effects of sulfonylurea herbicides on periphyton algae communities in stream water, periphyton from the Vemmenhog ¨ Stream was found to be much more tolerant to high concentrations of the sulfonylurea herbicide

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

tribenuron-methyl than periphyton collected from a nearby unpolluted stream ŽNystrom, ¨ 1997, p. 17.. In a report by Linders et al. Ž1994., providing an overview of pesticide data evaluated for registration purposes in Holland, the toxicity of 243 pesticides to water organisms was compiled. Data was available in this report for 31 of the 38 pesticides detected in stream water from the Vemmenhog ¨ catchment. Data were missing for the herbicides benazolin-ethylester, bromoxynil, clopyralid, flamprop-M, terbuthylazine and tribenuron-methyl and the metabolite desethylatrazine. Five pesticides detected in this study, the insecticide cyfluthrin and the herbicides atrazine, cyanazine, isoproturon and methabenzthiazuron exceeded during one or several weeks their respective toxicity value for the most sensitive species listed in the Dutch report. The maximum concentration detected of cyfluthrin, 5 m grl as an average concentration during 1 week in 1996, was 36 times higher than the LC 50 -value of 0.14 m grl Ži.e. the concentration required to kill 50% of the test organisms.. Cyfluthrin was detected on three other occasions at a weekly average concentration above its LC 50 -value. The maximum concentration detected for isoproturon, 10 m grl as an average concentration during a 2-week period in November 1996, was 16 times higher than the NOEC-value of 0.64 m grl Ži.e. the highest concentration without adverse effect on algae.. Altogether, isoproturon was detected at a weekly average concentration above its NOEC-value for 19 weeks. The corresponding figure for atrazine was 3 weeks, for cyanazine 2 weeks and for methabenzthiazuron 1 week. Another five herbicides Žmetamitron, metazachlor, prochloraz, prosulfocarb and simazine. had maximum, weekly average, concentrations just below their respective NOEC-values. For all other compounds detected concentrations were more than 10 times below the NOEC- or LC 50 -value given for the most sensitive species. The limit of determination ŽLOD. for most insecticides applied in the catchment was in the range 0.1]0.2 m grl. For a majority of the insecticides this LOD was close to, or even above, their respective LC 50 -value given in Linders et al.

247

Ž1994.. Only dimethoate, fenitrothion and pirimicarb had each an LOD value that was considerably below their respective LC 50 -value Ži.e. they had a better safety margin than the other insecticides.. This means that for several of the insecticides used in the catchment no definite statement can be made of their potential impact on the aquatic fauna in the stream. Pesticides detected in culvert discharge were often attenuated in the stream as a result of dilution by inflowing groundwater during low-flow situations and, for some of the pesticides, adsorption to sediments. However, it was also demonstrated that concentration peaks in the stream varied to a great extent, with daily average concentrations sometimes varying by one order of magnitude from one day to another. This means that concentrations entering the stream can, during short periods of time, be considerably higher, or lower, than those weekly average concentrations obtained by the sampling procedure most often used. The results from this investigation indicate that concentrations of some pesticides entering headwater streams in agricultural areas are close to, and during certain periods even above those levels demonstrated to have an impact on the aquatic flora and fauna. 5. Conclusion The overall findings of this investigation demonstrated that the occurrence of pesticides in surface water was a result of Ži. natural processes influenced by soil and weather conditions, together with the intrinsic properties of the compound, as well as Žii. point sources such as spills and non-agricultural application Že.g. in farmyards.. Some specific conclusions were as follows. v

v

v

Pesticide concentrations were higher during the application season, or shortly thereafter, with maximum concentrations occurring during runoff situations. Some pesticides were persistent in stream flow for several months after application and in some cases throughout the year. Some pesticides were detected at low concentrations for extended periods, regardless of

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

248

v

v

v

v

v

v

application period or stream flow. This could indicate that these pesticides are likely to persist in shallow groundwater. Based on the lower concentrations of pesticides detected in the open stream as compared to those in the culvert, it is apparent that wind drift had little influence on stream water quality in this catchment. For most pesticides there was a good correlation between amounts used and occurrence in the water samples. Losses of pesticides in catchment outflow, as a percentage of the applied amount, showed small variations between years and were independent of the pesticide application rate. High concentrations of pesticides were found in surface runoff collection wells in farmyards. This could have a potential impact on both stream and groundwater quality. Indeed, a substantial contribution of pesticide loss to stream water was from the application of pesticides in farmyards. Pesticides that occurred in the stream when there had been no preceding rainfall were presumed to be related to mismanagement in connection with filling and cleaning spraying equipment. Despite public concern and information campaigns directed towards those people applying pesticides during recent years, the relation between transported and applied amounts have been quite constant during the last 5 years. Lost amounts have, however, decreased in absolute figures, in accordance with de-

creased amounts applied during spring and early summer.

Acknowledgements Financial support for this work from the Swedish National Chemicals Inspectorate, the Swedish Environmental Protection Agency, the Environmental Foundation of the Malmohus ¨ county council and the County Government Board in Malmohus county is gratefully acknowledged. ¨ The pesticide analyses were performed by Malin ˚ kerblom-Branden, A ¨ ´ together with co-workers Eva ˚ Ramberg, Tomas Lundgren, Marit ¨ Peterson, Asa Berglof and Goran Jonsall, ¨ ¨ ¨ at the Department of Environmental Assessment, Organic Environmental Chemistry Section, Swedish University of Agricultural Sciences, Uppsala, Sweden. Investigation of tribenuron-methyl occurrence in water samples was performed in co-operation with Bo Nystrom ¨ and Hans Blanck, Botanical Institute, Department of Plant Physiology, Goteborg Uni¨ versity, with financial support from the Swedish National Chemicals Inspectorate. I would like to express my sincere gratitude to the farmers in the catchment for help and co-operation in this work; to Sten Hansson for his assistance with field work and expertise in the history of, and mechanisms operating in the catchment; to Goran Areskoug ¨ Žduring 1991]1996., Martin Larsson Ž1990. and Lennart Hansson Ž1990. for carrying out the interviews; and to Lars Bergstrom ¨ and Lennart Torstensson for help in reviewing this manuscript.

Appendix 1 Common and chemical names of pesticides included in the article Alpha-cypermethrin Atrazine Benazolin Bentazone Bromoxynil Chloridazon Clopyralid Cyanazine Cyfluthrin Cypermethrin 2,4-D Deltamethrin

w1 a Ž SU .,3 a x-Ž".-cyano-Ž3-phenoxyphenyl.methyl 3-Ž2,2-dichloroethenyl.2,2-dimethylcyclopropanecarboxylate 6-chloro-N-ethyl-N9-Ž1-methylethyl.-1,3,5-triazine-2,4-diamine 4-chloro-2-oxo-3Ž2 H .-benzothiazoleacetic acid 3-Ž1-methylethyl.-1 H-2,1,3-benzothiadiazin-4Ž3 H .-one 2,2-dioxide 3,5-dibromo-4-hydroxybenzonitrile 5-amino-4-chloro-2-phenyl-3Ž2 H .-pyridazinone 2,6-dichloro-2-pyridinecarboxylic acid 2-ww4-chloro-6-Žethylamino.-1,3,5-triazin-2-ylxaminox-2-methylpropanenitrile cyanoŽ4-fluoro-3-phenoxyphenyl.methyl 3-Ž2,2-dichloroethenyl.-2,2-dimethylcyclopropanecarboxylate cyanoŽ3-phenoxyphenyl.methyl 3-Ž2,2-dichloroethenyl.-2,2-dimethylcyclopropanecarboxylate Ž2,4-dichlorophenoxy.acetic acid w1 R-w1 a Ž SU .,3 a xx-cyanoŽ3-phenoxyphenyl.methyl 3-Ž2,2-dibromoethenyl.-

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

249

2,2-dimethylcyclopropanecarboxylate Dicamba Dichlobenil Dichlorprop Dichlorprop-P Diflufenican Dimethoate Diuron Esfenvalerate Ethofumesate Fenitrothion Fenpropimorph Fenvalerate Flamprop-M

3,6-dichloro-2-methoxybenzoic acid 2,6-dichlorobenzonitrile Ž".-2-Ž2,4-dichlorophenoxy.propanoic acid Žq.-2-Ž2,4-dichlorophenoxy.propanoic acid N-Ž2,4-difluorophenyl.-2-w3-Žtrifluoromethyl.phenoxyx-3-pyridinecarboxamide O,O-dimethyl S-w2-Žmethylamino.-2-oxoethylx phosphorodithioate N9-Ž3,4-dichlorophenyl.-N, N-dimethylurea w S-Ž RU , RU .x-cyanoŽ3-phenoxyphenyl.methyl 4-chloro-2-Ž1-methylethyl.benzeneacetate Ž".-2-ethoxy-2,3-dihydro-3,3-dimethyl-5-benzofuranyl methanesulfonate O,O-dimethyl O-Ž3-methyl-4-nitrophenyl. phosphorothioate cis-4-w3-w4-Ž1,1-dimethylethyl.phenylx-2-methylpropylx-2,6-dimethylmorpholine cyanoŽ3-phenoxyphenyl.methyl 4-chloro-a-Ž1-methylethyl.benzeneacetate N-benzoyl-N-Ž3-chloro-4-fluorophenyl.-D-alanine

Fluroxypyr

wŽ4-amino-3,5-dichloro-6-fluoro-2-pyridinyl.oxyxacetic acid 3-cyclohexyl-6-Ždimethylamino.-1-methyl-1,3,5-triazine-2,4Ž1 H,3H .-dione 4-hydroxy-3,5-diiodobenzonitrile N, N-dimethyl-N9-w4-Ž1-methylethyl.phenylxurea w1 a Ž SU .,3 a Ž Z .x-Ž".-cyanoŽ3-phenoxyphenyl.methyl 3-Ž2-chloro-3,3,3-trifluoro-1-propenyl.2,2-dimethylcyclopropanecarboxylate 3-cyclohexyl-6,7-dihydro-1 H-cyclopentapyrimidine-2,4Ž3 H,5H .-dione N9-Ž3,4-dichlorophenyl.-N-methoxy-N-methylurea Ž4-chloro-2-methylphenoxy.acetic acid Ž".-2-Ž4-chloro-2-methylphenoxy.propanoic acid Žq.-Ž R .-2-Ž4-chloro-2-methylphenoxy.propanoic acid 4-amino-3-methyl-6-phenyl-1,2,4-triazin-5Ž4 H .-one 2-chloro-N-Ž2,6-dimethylphenyl.-N-Ž1 H-pyrazol-1-ylmethyl.acetamide N-2-benzothiazolyl-N, N9-dimethylurea N-Ž1-ethylpropyl.-3,4-dimethyl-2,6-dinitrobenzenamine

Hexazinon Ioxynil Isoproturon Lambda-cyhalotrin Lenacil Linuron MCPA Mecoprop Mecoprop-P Metamitron Metazachlor Methabenzthiazuron Pendimethalin

Permethrin Phenmedipham Pirimicarb Prochloraz Propiconazole Propyzamide Prosulfocarb Simazine Terbuthylazine Triadimenol Tribenuron-methyl

Ž3-phenoxyphenyl.methyl 3-Ž2,2-dichloroethenyl.-2,2-dimethylcyclopropanecarboxylate 3-wŽmethoxycarbonyl.aminoxphenyl Ž3-methylphenyl.carbamate 2-Ždimethylamino.-5,6-dimethyl-4-pyrimidinyl dimethylcarbamate N-propyl-N-w2-Ž2,4,6-trichlorophenoxy.ethylx-1 H-imidazole-1-carboxamide 1-ww2-Ž2,4-dichlorophenyl.-4-propyl-1,3-dioxolan-2-ylxmethylx-1 H-1,2,4-triazole 3,5-dichloro-N-Ž1,1-dimethyl-2-propynyl.benzamide S-Žphenylmethyl. dipropylcarbamothioate 6-chloro-N, N9-diethyl-1,3,5-triazine-2,4-diamine 6-chloro-N-Ž1,1-dimethylethyl.-N9-ethyl-1,3,5-triazine-2,4-diamine b-Ž4-chlorophenoxy.-a-Ž1,1-dimethylethyl.-1 H-1,2,4-triazole-1-ethanol methyl 2-wwwwŽ4-methoxy-6-methyl-1,3,5-triazin-2-yl.methylaminoxcarbonylxaminoxsulfonylxbenzoate

References ˚ kerblom M, Thoren A av ´ L, Staffas A. Bestamning ¨ bekampningsmedel i dricksvatten. Var ¨ ˚ Foda ¨ Žin Swedish, with English abstract . 1990;42:236]243. Andersson A, Ohlin B. A capillary gas chromatographic multiresidue method for determination of pesticides in fruits and vegetables. Var ˚ Foda ¨ 1986;38:79]109. Baker DB, Richards RP. Herbicide concentration patterns in rivers draining intensively cultivated farmlands of Northwestern Ohio. In: Weigmann D, editor. Pesticides in terrestrial and aquatic environments. Blacksburg, VA, USA: Virginia Polytechnic Institute and State University, 1989: 103]120. Bergstrom ¨ LF, Jarvis NJ. Leaching of dichlorprop, bentazone, and 36 Cl in undisturbed field lysimeters of different agricultural soils. Weed Sci 1993;41:251]261.

Brown CD, Hodgkinson RA, Rose DA, Syers JK, Wilcockson SJ. Movement of pesticides to surface waters from a heavy clay soil. Pestic Sci 1995;43:131]140. Buhler DD, Randall GW, Koskinen WC, Wyse DL. Atrazine and alachlor losses from subsurface tile drainage of a clay loam soil. J Environ Qual 1993;22:583]588. Burgoa B, Wauchope RD. Pesticides in run-off and surface waters. In: Roberts TR, Kearney PC, editors. Environmental behaviour of agrochemicals. Chichester, UK: John Wiley, 1995:221]255. Clark L, Gomme J, Hennings S. Study of pesticides in waters from a chalk catchment, Cambridgeshire. Pestic Sci 1991;32:15]33. Daniel E. Beskrivning till jordartskartorna Tomelilla SV och Ystad NV. SGU Ae 99-100. Uppsala: Swedish Geological Survey, 1992:149 p. ŽIn Swedish, with English abstract..

250

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251

Frank R, Braun HE, van Hove Holdrinet M, Sirons GJ, Ripley BD. Agriculture and water quality in the Canadian Great Lakes Basin: V. Pesticide use in 11 agricultural watersheds and presence in stream water, 1975]1977. J Environ Qual 1982;11:497]505. Gaynor JD, MacTavish DC, Findlay WI. Surface and subsurface transport of atrazine and alachlor from a Brookston clay loam under continuous corn production. Arch Environ Contam Toxicol 1992;23:240]245. Gomme JW, Shurvell S, Hennings SM, Clark L. Hydrology of pesticides in a chalk catchment: surface waters. J Inst Water Environ Manag 1991;5:546]552. Griffini O, Bao ML, Barbieri C, Burrini D, Pantani F. Occurrence of pesticides in the Arno River and in potable water } a survey of the period 1992]1995. Bull Environ Contam Toxicol 1997;59:202]209. Harris GL, Bailey SW, Mason DJ. The determination of pesticide losses to water courses in an agricultural clay catchment with variable drainage and land management. B righ ton C rop P rotection C on feren ce-W eeds, 1991:1271]1278. Harris GL, Jones RL, Catt JA, Mason DJ, Arnold DJ. Influence of agricultural management and pesticide sorption on losses to surface waters. In: Walker A, Allen R, Bailey SW, Blair AM, Brown CD, Gunther P, Leake CR, Nicholls PH, ¨ editors. Pesticide movement to water, BCPC Monograph No 62. Farnham, UK: British Crop Protection Council, 1995:305]310. Harris GL, Nicholls PH, Bailey SW, Howse KR, Mason DJ. Factors influencing the loss of pesticides in drainage from a cracking clay soil. J Hydrol 1994;159:235]253. Hubbard RK, Williams RG, Erdman MD, Marti LR. Chemical transport from coastal plain soils under simulated rainfall: II. Movement of cyanazine, sulfometuron-methyl, and bromide. Trans ASAE 1989;32:1250]1257. Jaynes DB, Hatfield JL, Soenksen PJ. Water and chemical transport in surface and tile discharge of Walnut Creek. In: Toxic substances and the hydrologic sciences. American Institute of Hydrology, 1994:430]445. Johnson AC, Haria AH, Bhardwaj CL, Williams RJ, Walker A. Preferential flow pathways and their capacity to transport isoproturon in a structured clay soil. Pestic Sci 1996;48:225]237. Kirchmann H, Eriksson J. Properties and classification of soils of the Swedish long-term fertility experiments. II. Sites at ¨ ja and Orup. Acta Agric Scand Sect B, Soil Plant Sci O 1993;43:193]205. Kladivko EJ, Van Scoyoc GE, Monke EJ, Oates KM, Pask W. Pesticide and nutrient movement into subsurface tile drains on a silt loam soil in Indiana. J Environ Qual 1991;20: 264]270. Kreuger J. Provhanteringens paverkan pa ˚ ˚ halten av pesticider i vattenprov. Grant Report No 65422-8. Solna: Swedish Environmental Protection Agency, 1992:16 p. ŽIn Swedish.. Kreuger J. Study of pesticide stability during storage of water samples. Grant Report No 224 313. Solna: Swedish Environmental Protection Agency, 1994:7 p.

Kreuger J. Pesticides in stream water within a small catchment in southern Sweden. Report from the ‘Vemmenog¨ project’ 1990]1994. Technical Report No 32. Division of Water Quality Management, Swedish University of Agricultural Sciences, 1996:132 p. Kreuger J. Report from the ‘Vemmenhog-project’ 1995]1996. ¨ Pesticide concentrations and transport in water from a small agricultural catchment in southern Sweden. Technical Report No. 40. Division of Water Quality Management, Swedish University of Agricultural Sciences, 1997:57 p. Kreuger J, Tornqvist L. Multiple regression analysis of pesti¨ cide occurrence in streamflow related to pesticide properties and quantities applied. Chemosphere 1998; in press. Laroche A-M, Gallichand J. Analysis of pesticide residues in surface and groundwater of a small watershed. Trans ASAE 1995;38:1731]1736. Larson SJ, Capel PD, Goolsby DA, Zaugg SD, Sandstrom MW. Relations between pesticide use and riverine flux in the Mississippi River basin. Chemosphere 1995;31: 3305]3321. Larson SJ, Capel PD, Majewski MS. Pesticides in surface waters. Distribution, trends, and governing factors. In: Gilliom RJ, series editor. Pesticides in the hydrologic system, vol. 3. Chelsea, Michigan: Ann Arbor Press, 1997:373 p. Leonard RA. Movement of pesticides into surface waters. In: Cheng HH, editor. Pesticides in the soil environment: processes, impacts, and modeling, number 2. Madison, Wisconsin, USA: Soil Science Society of America, 1990: 303]349. Linders JBHJ, Jansma JW, Mensink BJWG, Otermann K. Pesticides: benefaction or Pandora’s box? A synopsis of the environmental aspects of 243 pesticides. Report No 679101014. The Netherlands: National Institute of Public Heath and the Environment ŽRIVM., 1994:204 p. Lundbergh I, Kreuger J, Johnson A. Pesticides in surface waters. A review of pesticide residues in surface waters in Nordic countries, Germany and the Netherlands and problems related to pesticide contamination. Council of Europe Press, ISBN 92-871-2776-X, 1995:54 p. Matthiessen P, Allchin C, Williams RJ, Bird SC, Brooke D, Glendinning PJ. The translocation of some herbicides between soil and water in a small catchment. J Inst Water Environ Manag 1992;6:496]504. Nystrom ¨ B. Metabolic indicators of ecotoxicological effects in freshwater periphyton communities. Thesis, Sweden: Botanical Institute, Plant Physiology, Goteborg University, ¨ 1997. Pereira WE, Rostad CE. Occurrence, distributions and transport of herbicides and their degradation products in the lower Mississippi River and its tributaries. Environ Sci Technol 1990;24:1400]1406. Richards RP, Baker DB. Pesticide concentration patterns in agricultural drainage networks in the Lake Erie basin. Environ Toxicol Chem 1993;12:13]26. Schottler SP, Eisenreich SJ, Capel PD. Atrazine, alachlor, and

J. Kreuger r The Science of the Total En¨ ironment 216 (1998) 227]251 cyanazine in a large agricultural river system. Environ Sci Technol 1994;28:1079]1089. Traub-Eberhard U, Kordel W, Klein W. Pesticide movement ¨ into subsurface drains on a loamy silt soil. Chemosphere 1994;28:273]284. Van der Linden AMA. Monitoring ground and surface waters: sampling strategy and implementation in legislation. In: Copin A, Houins G, Pussemier L, Salembier JF, editors. Environmental behaviour of pesticides and regulatory as-

251

pects. Proceedings of the 5th international workshop, COST 66. Brussels: European Study Service, 1994:299]305. Wauchope RD. The pesticide content of surface water draining from agricultural fields } a review. J Environ Qual 1978;7:459]472. Wauchope RD, Williams RG, Marti LR. Runoff of sulfometuron-methyl and cyanazine from small plots: effects of formulation and grass cover. J Environ Qual 1990;19: 119]125.