VäxtEko


Tidskrift/serie: Agraria
Utgivare: Sveriges lantbruksuniversitet (SLU)
Utgivningsår: 1999
Nr/avsnitt: 162
Författare: Kreuger J.
Adress: SLU, Department of Soil Sciences, P.O. Box 7072, SE-750 07 Uppsala, Sweden
Ingår i:
Titel: Pesticides in the environment - Atmospheric deposition and transport to surface waters
Huvudspråk: Engelska
Målgrupp: Forskare, rådgivare
Nummer (ISBN, ISSN): ISSN 1401-6249, IBN 91-576-5485-9

OBS! Fel i texten kan ha uppkommit då dokumentet överfördes från papper.

Pesticides in the environment - Atmospheric deposition and transport to surface waters

Jenny Kreuger

Department of Soil Sciences Uppsala

Akademisk avhandling som för vinnande av agronomic doktorsexamen kommer att offentligen försvaras i Undervisningshuset, sal L, SLU, Uppsala, fredag 28 maj 1999, kl. 10.00.

Abstract

This thesis deals with the environmental fate of pesticides, with special emphasis on occurrence and temporal and spatial distribution in atmospheric deposition, surface waters and sediments.

Pesticide residues in rainfall were investigated at two locations in the south of Sweden and one in the far north. There was a distinct geographic pattern with decreasing occurrence of pesticides in the remote areas of Sweden. Maximum concentrations occurred during the main application period May-June. Peak concentrations of 0.1-0.2 g/l were measured for several of the pesticides. The bulk deposition load varied between years, ranging from 3 to 24 g/m2 during the May-September period in southern Sweden.

Concentrations of pesticides in stream water from a catchment were observed throughout the sampling periods, with maximum concentrations occurring during application seasons and following run-off. Pesticides were also found in water as a result of spill. Peak concentrations of between 10 and 200 g/l were measured for several of the pesticides. Pesticide application for weed control in farmyards contributed to ∼ 20% of the overall pesticide load in stream water. Total amounts of pesticides lost in streamflow during May-September each year were generally less than 0.3% of the applied amount. The occurrence of pesticides in surface water was a result of natural processes influenced by soil and weather conditions, as well as point sources such as spills and non-agricultural application.

Pesticides were found in sediment samples, with a maximum concentration of 200 g/kg. Pesticides detected at the highest concentrations in sediment samples were either not detected or detected only at low levels in water samples collected in the stream during the same period. Pesticides with a log Pow-value above 3.5 were detected more frequently in the sediment samples compared to those with a lower log Pow-value.

Multiple linear regression analysis was used to express concentration, transported amount and loss rate in stream water as functions of different intrinsic pesticide properties (or a combination of these) and quantities applied. The results demonstrated that the quantity applied in the catchment area was the most important estimator of pesticide concentrations and transported amounts. Log Pow was the most significant intrinsic property for estimating relative loss of pesticides from the catchment.

Key words: pesticides, precipitation, rainfall, stream water, sediment, Sweden, catchment, water quality, pollution, monitoring, contamination, herbicides, fungicides, insecticides

Abstract

Kreuger, J. 1999. Pesticides in the Environment - Atmospheric Deposition and Transport to Surface Waters. Doctoral thesis. ISSN 1401-6249, ISBN 91-576-5485-9.

This thesis deals with the environmental fate of pesticides, with special emphasis on occurrence and temporal and spatial distribution in atmospheric deposition, surface waters and sediments.

Pesticide residues in rainfall were investigated at two locations in the south of Sweden and one in the far north. There was a distinct geographic pattern with decreasing occurrence of pesticides in the remote areas of Sweden. Maximum concentrations occurred during the main application period May-June. Peak concentrations of 0.1-0.2 g/1 were measured for several of the pesticides. The bulk deposition load varied between years, ranging from 3 to 24 g/m2 during the May-September period in southern Sweden.

Concentrations of pesticides in stream water from a catchment were observed throughout the sampling periods, with maximum concentrations occurring during application seasons and following run-off. Pesticides were also found in water as a result of spill. Peak concentrations of between 10 and 200 g/l were measured for several of the pesticides. Pesticide application for weed control in farmyards contributed to ∼ 20% of the overall pesticide load in stream water. Total amounts of pesticides lost in streamflow during May-September each year were generally less than 0.3% of the applied amount. The occurrence of pesticides in surface water was a result of natural processes influenced by soil and weather conditions, as well as point sources such as spills and non-agricultural application.

Pesticides were found in sediment samples, with a maximum concentration of 200 g/kg. Pesticides detected at the highest concentrations in sediment samples were either not detected or detected only at low levels in water samples collected in the stream during the same period. Pesticides with a log Pow-value above 3.5 were detected more frequently in the sediment samples compared to those with a lower log Pow-value.

Multiple linear regression analysis was used to express concentration, transported amount and loss rate in stream water as functions of different intrinsic pesticide properties (or a combination of these) and quantities applied. The results demonstrated that the quantity applied in the catchment area was the most important estimator of pesticide concentrations and transported amounts. Log Pow was the most significant intrinsic property for estimating relative loss of pesticides from the catchment.

The total deposition of the investigated pesticides corresponded to 30-240 mg/ha, which constitutes a small fraction of what is normally applied to an agricultural field. A very rough estimation gave an atmospheric deposition of ∼ 0.01% of the normal amount of pesticides applied. This figure was about one order of magnitude lower than the fraction leaving the catchment in streamflow.

Key words: pesticides, precipitation, rainfall, stream water, sediment, Sweden, catchment, water quality, pollution, monitoring, contamination, herbicides, fungicides, insecticides

Author's address: Jenny Kreuger, Swedish University of Agricultural Sciences, Department of Soil Sciences, P.O. Box 7072, SE-750 07 Uppsala, Sweden

Preface

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

  1. I. Kreuger, J. & Staffas, A. Atmospheric deposition of pesticides in rainfall in Sweden. Manuscript.
  2. II. Kreuger, J. 1998. Pesticides in stream water within an agricultural catchment in southern Sweden, 1990-1996. The Science of the Total Environment, 216: 227-251.
  3. III. Kreuger, J., Peterson, M. & Lundgren, E. 1999. Agricultural inputs of pesticide residues to stream and pond sediments in a small catchment in southern Sweden. Bulletin of Environmental Contamination and Toxicology, 62: 55-62.
  4. IV. Kreuger, J. & Törnqvist, L. 1998. Multiple regression analysis of pesticide occurrence in streamflow related to pesticide properties and quantities applied. Chemosphere, 37: 189-207.

The published papers are reproduced with the kind permission of Elsevier Science Ltd (Papers II and IV) and Springer-Verlag (Paper III).

Contents

Introduction

The commercial use of pesticides to control weeds and pests in agriculture developed soon after World War II. Pesticide use world-wide has increased dramatically since then and hundreds of different chemicals have been developed. Today the total agricultural use of pesticides in Europe is estimated to be about 350 000 tons of active ingredients (AIs) per year (SJV, 1996), which is about 28% of the total use throughout the world (Klassen, 1995).

During recent decades a growing concern about the environmental fate and impact of pesticide residues has emerged. Much of this concern has focused on the contamination of various water environments, due to the potential effects on water and sediment living organisms, and possible deterioration of irrigation and drinking water quality. Monitoring studies during the past decade have demonstrated the widespread occurrence of modem, agricultural pesticides in both groundwater and surface waters throughout Europe (Bester & Hühnerfuss, 1993; Albanis et al., 1994; Lundbergh et al., 1995; Mogensen & Spliid, 1995; Skark & Zullei-Seibert, 1995; Barcelo et al., 1996; Eke, 1996; Griffini et al., 1997; Garmouma et al., 1998; Spliid & Køppen, 1998), and also in Sweden (Kreuger & Brink, 1988; Åkerblom, 1991; Hessel et al., 1997; Rosling et al., 1998).

There are large regional differences within Europe regarding use patterns and amounts applied, e.g. half of the total European pesticide use is applied in only two countries, France and Italy, whereas the most intensive use (kg/ha) of pesticides is in the Netherlands (SJV, 1996). In Sweden some 90 different AIs are registered for use within the agricultural sector, with a total of ca. 1600 tons of AIs applied each year (Keml, 1998), half of which are applied in just one, densely populated county (Scania), in the far south of Sweden (SCB, 1997). Pesticides are also used in other sectors of society, e.g. industry, forestry, gardens and households. Moreover, accidental spills and dump sites provide still more environmental pesticide inputs. However, the greatest source of intentional environmental input world-wide is through agricultural inputs. In Sweden, agriculture is also a primary user of pesticides and due to the large acreage used for agricultural crops and the amounts involved, it was decided to focus the work presented in this thesis on the fate of pesticides in the environment as a result of agricultural input processes, with special emphasis on transport processes and occurrence in surface waters.

Objectives

During the late 1980's, discussion arose about the possible reasons for the frequent findings of pesticide residues in Swedish water courses. One source that had not been investigated was the possible contribution of atmospheric deposition of pesticides. To further explore the reasons for pesticide contamination in stream water, it was decided to go beyond the well-controlled conditions - either in the laboratory or using different sizes of lysimeters and field plots - under which, for good reasons, many environmental fate studies are done. The intention was to investigate pesticide sources, pathways and occurrence in an actual field situation. This was achieved by working within an agricultural catchment in close co-operation with the farmers operating within the catchment boundaries. The work was conducted within two major projects: (i) a rainwater investigation, that was planned and partly also carried out in parallel with similar studies in the other Nordic countries (Paper I), and (ii) a catchment monitoring study called the Vemmenhög-project (Papers II-IV).

Specific objectives were to:

Overview of pesticide transport processes

As illustrated in Fig. 1, there are a number of different transport processes that are important for the distribution of pesticide residues in the environment. The presence of agricultural pesticides in aquatic environments, which cannot be explained by point sources, is due to either atmospheric deposition (including drift) or transport from the field, as run-off or leaching. The following presentation will therefore give a brief overview of processes governing the transport of pesticides in the atmosphere and in the field. Description of processes involving transformation/degradation of pesticides in these environmental compartments, although important, is beyond the scope of this overview and is therefore not covered specifically.

Pesticides in the atmosphere

Atmospheric transport and subsequent deposition is an important pathway for the distribution of pesticides to land and water located both close to and far from the source areas. Once released into the atmosphere pesticides may undergo a variety of physical and chemical reactions. Many pesticides are semivolatile, which means that they are transported in the atmosphere both in the gas and particle phases and the distribution of the pesticides between the phases is important for the deposition process. The behaviour of pesticide residues in the atmosphere has most likely been the least studied area of pesticide fate in the environment (Roberts, 1996).

Fig. 1. Principal environmental pathways by which agricultural pesticides may be transported to surface waters.

Input to the atmosphere

During the application process considerable quantities of the sprayed pesticide fail to reach the soil or plant target and are thus lost to the atmosphere as spray drift. The magnitude of drift losses is largely dependent on the type of compound formulation, application methodology and local meteorology, such as wind speed and temperature (Elliott & Wilson, 1983). There are difficulties involved in estimating the total losses resulting from off-target drift as a percentage of the amount applied, partly due to the complexity of the different processes involved and also to the lack of standardised methods for measuring these losses (Arvidsson, 1997). Therefore, drift losses have, to a large extent, been documented and visualised in the past by damage to sensitive crops at some distance from the site of application (Hagenvall, 1990). Nevertheless, there have been investigations that have documented drift losses during application to range from 1 to 75% of the applied amount (Majewski & Capel, 1995; Arvidsson, 1997).

Losses to the atmosphere also occur after application through volatilisation and wind erosion. These processes depend primarily on climate, the type of active ingredient, formulation and characteristics of soil and plants at the target site (Plimmer, 1988). Volatilisation (further described in a later section in this overview) is often considered to be the most important process for losses of a pesticide to the atmosphere (Taylor, 1995). It is a process that to a large extent is governed by the Henry's Law constant (HLC) of the pesticide, i.e. the ratio of air-water partitioning based on the compound vapour pressure and water solubility (Sunito et al., 1988). Depending on the volatility of the compound, application method and soil moisture, the mass of applied pesticide lost to the atmosphere by volatilisation has been found to range from 2-90% (Majewski & Capel, 1995).

However, as described above, there are other ways for pesticides to enter into the atmosphere that are independent of the intrinsic properties of the compound. It is therefore likely that a certain amount of most pesticides applied in the field enter into the atmosphere either during or soon after the application process.

Transport in the atmosphere

Once a pesticide becomes suspended in the atmosphere it will distribute itself between the vapour, aqueous and particle phases in order to reach an equilibrium state. This is shown schematically in Fig. 2. The distribution of a pesticide between these phases is dependent on the physical and chemical properties of the compound, such as water solubility and vapour pressure, as well as environmental factors such as temperature, moisture and the nature and concentration of suspended particulate matter (Majewski & Capel, 1995). Most high-molecular-weight organic compounds lie between the extremes of being either only in the vapour, or only in the particulate phase, and their distribution and atmospheric lifetimes depend to a large extent on the particle concentration and composition (e.g. size, surface area and organic carbon content) in the atmosphere (Eisenreich et al., 1981). Partitioning of pesticides to the particle phase in the atmosphere is favoured by lower temperatures (Atkinson et al., 1992).

Movement of pesticides in the atmosphere takes place through dispersion, which is a combination of diffusion and transport processes that occur simultaneously (Schroeder & Lane, 1988). Diffusion, which promotes the dispersion of gases and atmospheric particles (aerosols), is caused by turbulent motions that develop in air that is unstable. Transport, on the other hand, results from air-mass circulation driven by local or global forces. Certain meteorological conditions such as thunderstorms can move these airborne pesticide vapours and particles into the upper troposphere. Once there, they can be distributed regionally and even globally (Majewski & Capel, 1995). The actual distance travelled by pollutants strongly depends on the amount of time a specific pollutant resides in the atmosphere and is available for dispersion.

Fig. 2. Outline of possible atmospheric partitioning pathways (Majewski, 1991).

When pesticides are in the gas phase they are more likely to undergo photolysis and different kinds of chemical reactions, e.g. with OH-radicals, NO3-radicals and 03, and may have lifetimes in the order of a few hours or less. The potential impact of such pesticides is then restricted to local areas around the point of application. The reaction with the OH-radical is in general the dominant process for the majority of organic pollutants (Van Pul et aL, 1998). Chemical reactions on particles and in the liquid phase (cloud water) are considered to be of little importance. When pesticides are attached to particulate matter they are instead more susceptible to physical loss processes of wet and dry deposition of the particles. The distribution of pesticides between gas, particle and water phases dictates the atmospheric lifetime of the pesticide through effects on the transformation reaction rates and depositional rates to a surface.

Deposition

Airborne pesticides can be removed from the atmosphere and deposited on land and water surfaces by (i) wet deposition, (ii) dry particle deposition and (iii) gas-exchange at the air-surface interface (Fig. 2).

Wet deposition occurs when pollutants are washed out of the atmosphere by rain and snow and includes both gaseous and particle associated compounds. However, different removal mechanisms apply depending on whether the compound is in the gaseous phase or particle associated. Wet deposition of gases via rain (vapour wash-out) results from the dissolution of vapours into raindrops and is governed by the HLC, which dictates the rate by which a gas reaches equilibrium with a raindrop (Eisenreich et al., 1992). Wet deposition of particulate matter via rain (particle wash-out) occurs either as water vapour condenses around a particle during the formation of a raindrop (incorporation) which then falls to the soil/water surface, or as a raindrop physically impacts a particle as it falls (collection) (Majewski & Capel, 1995).

Dry deposition is a continuous process and occurs when particles are deposited on surfaces when there is no precipitation. The deposition rate is dependent on several factors, such as the size of the particle, the deposition layer and macro- and micrometeorology (Eisenreich et al., 1992). The size of the atmospheric particle is the chief determinant of the deposition mechanism and rate. Dry deposition is a complex process, mainly because atmospheric particles occur within a wide range of sizes, with differences in surface area and mass of the particle. However, although larger particles usually weigh more than smaller ones and therefore tend to settle faster, most of the sorbed pesticide may be concentrated on the smaller particles because of the higher surface area-to-volume ratio (Majewski & Capel, 1995). At greater distances from areas of application, dry particle deposition is usually less important than wet deposition of fine particles (Eisenreich et al., 1992).

Gas-exchange is a two-way process involving uptake and revolatilisation of atmospheric vapours from land and water surfaces. Generally, the mass transfer is considered to be limited by the rate of molecular diffusion through thin films of air and water on either side of the surface (Bidleman & McConnell, 1995). The net flux is driven by a deviation from chemical equilibrium between the air and water phase. The gas-exchange-process can be of considerable importance for fluxes of pollutants to oceans and large lakes. For certain persistent compounds it is considered to be of even greater magnitude than fluxes due to precipitation and dry particle deposition (Bidleman & McConnell, 1995).

The relative importance of the different deposition processes is to a large extent unclear. This is partly due to the difficulties in directly measuring dry deposition and gas-exchange rates, which both have a high degree of uncertainty associated with them.

Pesticides in the agricultural field

After the application of a pesticide to an agricultural field, its continuing fate is determined by numerous biological, physical and chemical processes that can either promote or prevent pesticide transport away from the site of application. Transport of pesticides from the field to surface water bodies occurs primarily by means of volatilisation, run-off or leaching. The principal factors affecting transport include: (i) inherent properties of the compound such as water solubility, vapour pressure and octanol/water partition coefficient; (ii) environmental conditions, including climate; (iii) soil properties, such as soil texture and organic matter content; (iv) landscape characteristics including topography; and (v) management practices, such as tillage, crop selection and application methodology. Leaching of pesticides can, for example, be promoted by certain combinations of pesticide and soil properties (e.g. low soil adsorption potential of the pesticide and high soil infiltration rate), which increase the chances of movement to tile drains and groundwater, but decrease run-off potential. On the other hand, when the physical characteristics of the site promote run-off (e.g. soils with low infiltration capacity and sloping topography), overland surface transport of the pesticide becomes more likely.

Volatilisation

Volatilisation is often one of the principal pathways by which pesticides are lost from soils and crops after application in the field, although losses also occur as wind erosion of particulates from the soil surface. A volatilising pesticide will first diffuse through a stagnant air boundary layer immediately above the treated surface and then be carried away by turbulent flow of air (Seiber & Woodrow, 1995). In the soil, pesticides can be transported upward, to the soil surface, either by gaseous diffusion through the airfilled volume of the soil, or by the upward flow of soil solution induced by water evaporation. In either case, the concentration in soil-air at the surface will be governed by the soil water-air equilibrium and the pesticide's HLC-value (Sunito et al., 1988). Pesticides with high HLCs (i.e. high vapour pressure and/or low water solubility) are those most likely to be lost to the atmosphere. The rate is also influenced by the movement of air over the surface. Other important factors controlling the rate of volatilisation are the method of application, the soil moisture distribution, soil organic matter content, soil temperature and soil tillage practices (Majewski, 1991).

Taylor (1995) points out that volatilisation losses often exceed 10% of the applied amount within 48 hours or less. In rare cases, as much as 90% of the applied amount can be lost when residues of volatile pesticides are exposed on moist soil or plant surfaces. However, in a dry soil, the rate of evaporation may be low, especially if the compound is strongly adsorbed. When the soil is wetted, the pesticide may be displaced from adsorption sites and again subjected to volatilisation.

Run-off

Leonard (1990) defined run-off as water and any dissolved or suspended matter it contains that leaves a field in surface drainage. He concluded that, at the fieldscale, direct surface run-off (overland flow) is the major component of run-off, whereas interflow (i.e. return of subsurface water to the soil surface by seepage) at this scale is minimal. Run-off from a field begins when rainfall or irrigation rates exceed infiltration rates. The pesticide run-off potential is primarily affected by factors that influence the persistence of the pesticide in the run-off active zone at the soil surface (i.e. the pesticide's potential for degradation, volatilisation and leaching below the soil surface). Run-off concentrations will decrease exponentially as the time between application and a run-offgenerating storm increases, with half-lives of between several days to 2 weeks. For this reason the majority of pesticide losses during a season, resulting from run-off, are usually due to a single storm event occurring soon after application (Wauchope & Leonard, 1980).

During run-off, the pesticide may be transported either dissolved in water or bound to soil particles, or both. Leonard (1990) suggests that "at the field microscale, pesticide extraction into run-off may be described as mechanisms of (i) diffusion and turbulent transport of dissolved pesticide from soil pores to the run-off stream; (ii) desorption from soil particles into the moving liquid boundary; (iii) dissolution of stationary pesticide particulates; or (iv) scouring of pesticide particulates and their subsequent dissolution in the moving water. Pesticides are also entrained in run-off attached to suspended soil particles".

For pesticides that are not strongly bound to soil, e.g. pesticides with a water solubility above ca. 10 mg/l, most of the transported mass will be in the water phase of the run-off. However, insoluble, hydrophobic pesticides with a solubility of less than 1 mg/l will be tightly bound to soil and most or all loss will occur in the sediment phase (Wauchope et al., 1995). Nevertheless, most pesticides are lost in the water phase, simply because sediment is usually only a small fraction, by weight or volume, of run-off. The result of this is that erosion-control practices, except when they control water as well as sediment losses, can be expected to have little effect on run-off losses of pesticides which have solubilities exceeding 1 mg/l (Wauchope, 1978).

Burgoa & Wauchope (1995) summarised the general features describing run-off losses of pesticides from the field: (i) run-off losses are generally 0.5% or less of applied amounts, although losses of up to 5% or even higher are possible under worst-case conditions; (ii) storm timing is a critical determinant of pesticide run-off losses; (iii) application target (e.g. soil or foliage) and compound formulation are important factors; and (iv) the majority of pesticides are lost in the water phase of run-off.

Leaching

The transport of pesticides in the soil to drainage tiles and groundwater is largely controlled by the movement of water. Water flow through soils is governed by the driving force (hydraulic gradient) and the hydraulic conductivity of the soil. These forces, described in numerous articles and summarised by e.g. Rao et al. (1988), determine the rate with which water moves within the soil profile. The conductivity varies according to soil type and is primarily regulated by the size, shape and interrelation of the soil pores. In the unsaturated zone above the water-table, a gas phase in the soil also controls the flow paths of the water, with flow rates thus being determined by the water content of the soil as well. Since the water content varies considerably within the unsaturated zone (Dekker et al., 1999), this results in water flow rates above the water-table being more spatially variable and hence less easily predicted than those within the saturated zone (Barbash & Resek, 1996).

The influence of differences in soil types on water flow rates can be illustrated by an example given by Brown et al. (1995a): in a sandy soil, flow is very rapid in the saturated state, but quickly diminishes as soil suction increases, i.e. when the larger pores have drained. In a clayey soil matrix, flow is slower in the saturated state, but faster in the unsaturated state because there are a greater number of finer soil pores contributing to flow than in the sandy soil at the same suction.

The flow of water through the porous structure of soils is normally slow enough for soil sorption, for solute exchange between regions of mobile and immobile water and for degradation processes to influence the movement of the pesticide through the soil profile (Bergström & Stenström, 1998). However, water can also be transported rapidly through preferential flow to considerable depths beneath the soil surface, with only a small fraction of the available pore space being used to move water and solutes. Preferential flow is not restricted to 'cracking' clay soils, but, as pointed out by e.g. Hallberg (1989), has been shown to occur in soils of all textures, from clays, to silt loams and even sands. Three categories of preferential flow, with underlying different physical mechanisms, have been suggested (Gish et al., 1998; Jarvis, 1998): (i) macropore flow, i.e. flow through large, continuous, structural pores such as soil cracks, worm holes and root channels; (ii) fingering, i.e. flow resulting from the formation of an unstable wetting front, which mainly occurs in more homogeneous coarse-texured soils; and (iii) funnel flow, i.e. flow in heterogeneous soils containing soil lenses and mixtures of various particle size fractions. The importance of leaching processes, with special emphasis on preferential flow, for the movement of pesticides in soil profiles was recently reviewed by Larsson (1999).

Although leaching of pesticides in the soil to a large extent is controlled by the movement of water, it is also, naturally, governed by the intrinsic properties of the pesticide (notably its soil dissipation half-life and soil organic-carbon partition coefficient), soil characteristics and management practices. The influence of different combinations of these features has been documented in numerous studies, performed both in the laboratory and in different sized lysimeters and field plots. Flury (1996) summarised a number of studies reporting pesticide losses through experimental tile-drained fields and concluded that leaching losses were most often in the range < 0.1 to 1% of applied amounts, with occasional losses reaching 4%.

Tile drainage water

Water usually enters a natural stream through surface run-off and through subsurface flow. However, if the water table in the area draining to the stream has been manipulated by the installation of tile-drainage systems (see Paper II), most of the subsurface flow will take place through the drainage tiles. Also, the run-off potential probably decreases if tile drains are installed since they lower the water-table and enhance the infiltration rate. However, the water leaving the field by means of tile drains can result from both subsurface flow and run-off, if surface water inlets are connected to the tile-drainage system. Larson et al. (1997) stated that run-off occurs by overland flow, interflow and flow through tile-drainage networks. Schottler et al. (1994), investigating pesticide transport to surface water in an area with complex subsurface tile-drainage networks, considered three major pathways for water to reach a stream or a river: (i) a surface transport pathway, including both overland flow and removal of ponded water via vertical tile drains; (ii) a subsurface flow pathway with water that travels through the unsaturated zone and is collected and transported by the tile lines; and (iii) an additional subsurface pathway with water that infiltrates to the water-table and is thereafter moved to a river in the local and/or regional groundwater system. Clearly, water enters drainage tiles as a result of different flow processes, i.e. both run-off and leaching. However, it must be remembered that the mere presence of tile drains increases the infiltration rate and decreases run-off potential, which means that a larger part of the tile-drainage water would, in the normal field situation when the soil is unfrozen, result from leaching through the unsaturated zone of the soil profile.

Methods

Determination of atmospheric deposition of pesticides

Proper assessment of the atmospheric deposition load of pesticides includes measurements of wet deposition, dry particle deposition and gas-exchange at the air-surface interface, as illustrated in Fig. 2. However, as stated by e.g. Eisenreich et al. (1992) and Lee & Nicholson (1994), the latter two processes, which constitute dry deposition, are difficult to assess and quantify properly using current sampling methodology. Wet deposition (i.e. precipitation as rain or snow), on the other hand, may be measured on a routine basis and is commonly collected using different kinds of wet-only collectors (Strachan & Huneault, 1984). These samplers are covered and only open when it rains. To distinguish between particle bound pesticides and pesticides dissolved in the precipitation, the water should be filtered as it is collected. However, the value of this information depends on the assumption that no exchange between dissolved and particulate phases occurs within precipitation or during storage (Lee & Nicholson, 1994).

Wet deposition is also frequently collected using bulk (passive) samplers, open to the atmosphere all the time. Bulk samplers are often easy to handle and less complicated than the wet-only collectors and give an idea of total deposition resulting from both wet and dry deposition. However, such measurements do not explain the wash-out mechanisms. Nevertheless, the choice of the type of deposition sampler must depend on the purpose of the measurements. The study reported in Paper I aimed at estimating the amounts of selected pesticides deposited by atmospheric transport in relation to those entering stream water through run-off and subsurface flow. To achieve this objective, it was only necessary to know bulk deposition, so the sampling strategy was designed accordingly.

Precipitation samplers may vary in size, sampling characteristics and surface characteristics. It is important that the material in the sampler is inert and non-contaminating. In a comparison of samplers with different surface materials (Teflon and stainless steel) and surface area of the collector, Eisenreich (1990) found little difference between the samplers when collecting organic pollutants such as PCBs and PAHs. The size of the collection area of the runnels determines the volume of water collected for each rain event. Therefore, the size of the runnel used will indirectly influence the level of detection for the pesticides included in the investigation, since the analytical detection limit of pesticides is dependent on the volumes of water extracted. This can be overcome by, for example, using long sampling intervals or by having several parallel funnels. However, long sampling intervals increase the risk of re-emission of deposited pesticides, as well as possible losses due to transformation reactions while the sample is left unattended in the field.

In this study, bulk samplers with either 0.03-0.15 m2 glass runnels or 0.5-1.0 m2 stainless steel funnels were used (Paper I). Rainwater was collected in large glass bottles, with a preservative (dichloromethane) added to prevent degradation losses and adsorption of hydrophobic compounds to the glass sides. Rainwater collected with the stainless steel runnels were stored at 4C in a refrigerator placed under the funnel during the collection period, in order to minimise losses due to degradation and evaporation, both processes promoted by higher temperatures. A comparison at one of the sites (Lurbo) demonstrated a good agreement between the different samplers used, with recoveries between parallel samplers within the normal variance of the analytical procedures.

Determination of pesticide occurrence in stream water

There arc many different sampling strategies for the collection of stream water for pesticide monitoring purposes. However, the selection of a sampling strategy is, as always, a compromise between the purpose of the study and the resources available.

There are basically two types of samples that can be collected from a water source (Dick, 1994): discrete samples or composite samples. A discrete sample (grab sample) is an individual sample collected at a fixed time and deposited in its own individual container. It represents the source at the time the sample was collected. A composite sample consists of two or more smaller samples collected at different times and deposited into the same container. It represents the average characteristics of the source during the period when the sub-samples were collected.

Depending on the objective, the collection of multiple samples can be performed either at equal time increments or in proportion to the water flow. A time-paced composite sample consists of sub-samples of equal volume collected at equal time intervals and deposited in the same container. A flow-proportional composite sample consists of sub-samples of equal volume collected at equal discharge intervals. A flow-proportional composite sample can also consist of sub-samples with volumes proportional to the flow volume and collected at equal time intervals (Dick, 1994).

Surface water samples may be collected manually by simply submerging a sampling device below the water surface. However, to facilitate the sampling procedure, automatic sampling devices are frequently used in monitoring studies today. These samplers can be automatically controlled to collect multiple discrete samples or composite samples, either time-paced or flow-proportional.

Pesticide concentrations can increase rapidly in stream waters following pesticide applications and intense rainfall (Richards & Baker, 1993). In such cases, sampling strategies can be targeted towards frequent sampling during storm hydrographs, with less sampling required between storms. In contrast, the sampling procedure for an irrigated agricultural area will have to adopt another strategy, since pesticides in such areas are transported to surface waters by drainage from irrigated fields with little possibility to predict when high pesticide concentrations may occur (Domagalski, 1997). Furthermore, pesticide residues resulting from wind drift or improper handling and other forms of spill may enter the stream regardless of stream hydrograph and, in such cases, it is difficult to determine when peak concentrations or loadings may occur.

The size of the area draining to the sampling site is important, since the magnitude and duration of concentration peaks are correlated to the drainage area. As demonstrated by Richards & Baker (1993), pesticide concentrations, in general, increase as catchment size decreases. The probability of sampling at or very near the time of peak concentration decreases as the stream size decreases, given a fixed frequency sampling program, due to the shorter run-off hydrographs.

Richards (1993) summarised the difference between the main sampling strategies in the following way: a flow-weighted concentration takes account of the fact that more water passes by the station during high flow than during low flow and is therefore appropriate for estimating loading rates. A time-weighted concentration ignores the amount of water passing the station and is instead more appropriate for questions of exposure to animals and plants living in the water or to humans via drinking water abstraction (which usually takes place at a more or less constant rate).

Often, for management decisions, the evaluation of a monitoring programme is based on the frequency at which concentrations are detected above a regulatory level. In such cases the sampling interval is of great importance. For example, Domagalski (1997) showed that the concentration level set for protection of aquatic life of an insecticide, was exceeded in 40% of the samples when collecting grab samples three times a week, but only in 25% of the samples when collecting one sample a week. He concluded that a time-weighted sample, collected with an auto-sampler, might be the best option.

In this study, automatic water samplers were used (Paper II). The samplers were programmed to collect time-paced composite samples with collection intervals ranging from daily to weekly and with sub-samples taken at 10-min or hourly intervals. During the initial phase of the monitoring project discrete (grab) samples were also collected in different parts of the catchment, as well as in parallel with the automatic samplers for comparison purposes. Comparison between composite samples and grab samples collected around twice a week showed that the discrete samples usually underestimated the pesticide load. However, a pesticide was sometimes only detected in an occasional grab sample, and not in the composite sample collected at that time. This was possibly the result of a small peak being rapidly transported past the point of collection, with the result that these compounds were attenuated in the composite sample to a concentration below the limit of detection and therefore undetected in the composite samples. Collection of daily, and a few hourly, composite samples demonstrated rapid fluctuations of pesticide concentrations detected in streamflow, sometimes varying by one order of magnitude from one day to another. The fluctuations were most often due to changes in streamflow, but peak concentrations were also found to occur without any changes in streamflow.

Summary and reflections on results presented in Papers I-IV

Precipitation

Pesticides were monitored in rainfall at two locations in the south of Sweden and one in the far north (Paper I) to examine the occurrence and temporal and spatial distribution of selected pesticides in atmospheric deposition. Seventeen pesticides and one metabolite were detected in rainfall from southern Sweden, with HCHs, phenoxy acids and triazines being those most commonly detected. In rainwater from the far north of Sweden, low concentrations of HCHs were detected, with traces of phenoxy acids and atrazine occurring on single occasions. This indicates a distinct geographic pattern with decreasing occurrence and concentrations of pesticides in the remote areas of Sweden. Maximum concentrations of the pesticides in rain occurred during the main application period May-June and then they declined by late summer. However, some pesticides were present at measurable levels until the end of the sampling period in late September. Peak concentrations of 0.1-0.2 g/1 were measured for several of the pesticides in individual samples. Most samples contained a number of different pesticides, with up to 13 different pesticides detected in a single sample during the main application season. The bulk deposition load varied between years, ranging from 3 to 24 g/m2 during the May-September period in southern Sweden.

Based on the results presented in Paper I, it can be concluded that pesticides in current use are widespread in atmospheric deposition in Sweden, along with others such as 2,4-D, atrazine and HCHs that are no longer registered for use within the country. This indicates a transboundary atmospheric transport and deposition of airborne organic pollutants in Swedish rainfall.

Surface water

Water samples collected from streamflow in the Vemmenhög catchment (Paper II) were analysed for a wide range of pesticides applied within the catchment. They were examined for occurrence and temporal variability of pesticide residues, and these concentrations were related to patterns of use and transport pathways within the catchment.

Thirty-seven pesticides and one metabolite were detected in streamflow leaving the catchment, with a range of different herbicides being those most commonly detected. Concentrations of pesticides in stream water were observed throughout the sampling periods, with maximum concentrations occurring during application seasons and following run-off situations. Pesticides were also found in water samples as a result of careless handling and application procedures. Peak concentrations of between 10 and 200 g/1 were measured for several of the pesticides with concentrations sometimes varying by one order of magnitude from one day to another. Some pesticides were detected at low concentrations for extended periods, regardless of application period or streamflow. This could indicate that these pesticides are also likely to persist in shallow groundwater.

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. Based on the lower concentrations of pesticides detected in the open stream as compared to those in the culvert, it was apparent that wind-drift had little influence on pesticide concentration in this catchment. 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. Pesticide application for weed control on farmyards contributed to ∼ 20% of the overall pesticide load in stream water. Pesticides occurred in the discharge throughout the winter and originated from both autumn and previous spring applications, as well as from farmyard application. Some autumn applied pesticides prevailed in streamflow during the following summer. Total amounts of pesticides lost in streamflow during May-September each year varied between 0.5 and 2.8 kg during the seven-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 and were independent of the pesticide application rate.

Table 1. Estimated loss of pesticides as a percentage of amounts applied in drainage area
Pesticide Size of study site (km2) Loss (%) Reference
Catchments      
alachlor 69,300-2,914,000 0.10-0.47 (Battaglin et al., 1993)
alachlor 38,585 0.04-0.20 (Schottler et al., 1994)
atrazine 19-79 0.96 (Frank et al., 1982)
atrazine ∼ 3.000,000 0.4-1.7 (Pereira & Rostad, 1990)
atrazine 3.998-6,840 0.82-1.96 (Frank et al., 1991c)
atrazine 160 1.52 (Gomme et al., 1991)
atrazine 69,300-2,914,000 0.58-1.83 (Battaglin et al., 1993)
atrazine 935 0.16-1.81 (Albanis et al., 1994)
atrazine 27 0.23 (Jaynes et al., 1994)
atrazine 38,585 0.33-0.62 (Schottler et al., 1994)
atrazine ∼ 30,000-3,000,000 0.62-1.9 (Larson et al., 1995)
atrazine 35 0.15 (Ng & Clegg, 1997)
atrazine 51 0.18-5.6 (Jaynes et al., 1999)
chlortoluron 160 0.08 (Gomme et al., 1991)
cyanazine 69 300-2 914 000 0.56-2.77 (Battaglin et al., 1993)
cyanazine 38,585 0.32-1.30 (Schottler et al., 1994)
cyanazine ∼ 30.000-3,000,000 0.57-3.1 (Larson et al., 1995)
2,4-D 19-79 0.15 (Frank et al., 1982)
2,4-D 935 0.07-0.69 (Albanis et al., 1994)
dichlorprop 9 0.01-0.56 Paper II
ethofumesate 9 0.02-0.24 Paper II
fenpropimorph 9 0.004-0.03 Paper II
isoproturon 160 0.10 (Gomme et al., 1991)
MCPA 19-79 0.05 (Frank et al., 1982)
MCPA 935 0.07-0.14 (Albanis et al., 1994)
MCPA 9 0.06-0.21 Paper II
mecoprop 160 0.05 (Gomme et al., 1991)
mecoprop 9 0.07-0.44 Paper II
metamitron 9 0.01-0.24 Paper II
metolachlor 3,998-6,840 0.002-1.16 (Frank et al., 1991c)
metolachlor 69.300-2,914,000 0.50-1.07 (Battaglin et al., 1993)
metolachlor ∼ 30.000-3,000,000 0.45-0.93 (Larson et al., 1995)
metolachlor 35 0.10 (Ng & Clegg, 1997)
metolachlor 51 0.05-1.6 (Jaynes et al., 1999)
pirimicarb 9 0.01-0.16 Paper II
propiconazole 9 0.02-0.44 Paper II
propyzamide 160 0.38 (Gomme et al., 1991)
simazine 19-79 0.07 (Frank et al., 1982)
simazine 160 0.64 (Gomme et al., 1991)
simazine ∼ 30,000-3,000,000 0.57-10.5 (Larson et al., 1995)
tri-allate 160 0.02 (Gomme et al., 1991)
Field sites      
atrazine 5 ha 0.15 (Muir & Baker, 1976)
atrazine 5 ha 0.05 (Jaynes et al., 1994)
atrazine 14 ha 0.09-1.9 (Frank etal., 199 la)
isoproturon 5 ha 0.40 (Traub-Eberhard et al., 1994)
isoproturon 1.1 ha 0.09 (Traub-Eberhard et al., 1994)
metolachlor 14 ha 0.003-0.01 (Frank et al., 1991b)
pendimethalin 5 ha < 0.001 (Traub-Eberhard et al., 1994)
Field plots      
alachlor 0.1-0.4 ha 0.00-0.01 (Kladivko etal., 1991)
atrazine 0.02 ha 0.009-0.13 (Buhler etal., 1993)
atrazine * ≤ 1.2 (Johnson et al., l995b)
atrazine 0.1-0.4 ha 0.01-0.06 (Kladivko et al., 1991)
carbofuran 0.1-0.4 ha 0.05-0.94 (Kladivko et al., 1991)
cyanazine 0.1-0.4 ha 0.00-0.04 (Kladivko et al., 1991)
isoproturon 0.25 ha 0-0.45 (Brown et al., l995b)
isoproturon 0.2 ha 0.5 (Harris et al., 1994)
isoproturon * 1.7-3.3 (Harris etal., 1995)
isoproturon 0.1 ha 1 (Johnson et al., 1995a)
mecoprop 0.25 ha 0-0.04 (Brown et al., 1995b)
pendimethalin * 0.009-0.04 (Hams et al., 1995)
prochloraz * 0.02-0.05 (Harris et al., 1995)
triasulfuron * 2,9-5.2 (Harris et al., 1995)
trifluralin 0.25 ha 0.001-0.02 (Brown et al., l995b)

* = Size of field plots not reported.

Based on the results presented in Paper II it can be concluded that the occurrence of pesticides in surface water was a result of (i) natural processes influenced by soil and weather conditions, together with the intrinsic properties of the compound, as well as (ii) point sources such as spills and non-agricultural application (e.g. in farmyards).

Pesticide losses, expressed as a percentage of applied amounts, have been estimated in a number of different studies which are summarised in Table 1, with results from Paper II included. The size of the study areas varied from small-scale plots of < 1 ha to large, heterogeneous catchments of thousands of km2. For a majority of the pesticides investigated, total losses to surface waters were 0.5% or less, although losses of up to 10% were recorded. Estimated losses most often varied within one or two orders of magnitude, quite independently of the pesticide used and the size of the study area, although there was a tendency for losses found at the catchment scale to be slightly larger than those found under more controlled conditions at the field scale. This indicates a contribution from point sources and, also, uses in urban areas not included in the estimations of applied amounts in the larger catchments.

Sediments

Stream and pond sediment samples collected in the Vemmenhög catchment (Paper III) were analysed for a range of pesticides to determine concentration levels and to relate findings of these pesticides to those occurring in stream water from the same area. Eleven pesticides were detected in sediment samples from the catchment, with fungicides and insecticides being those most commonly detected. The maximum concentration detected was 200 g/kg for a single pesticide. Pesticides detected at the highest concentrations in sediment samples were either not detected or detected only at low levels in water samples collected in the stream during the same period. Pesticides with a log Pow-value above 3.5 were detected more frequently in the sediment samples compared to those with a lower log Pow-value, even though the former were either not used at all or their use was limited during the investigation period. Pesticide residue analysis of sediment samples demonstrated the importance of pesticide distribution between matrices.

Factors controlling pesticide occurrence

Because pesticides applied in the Vemmenhög catchment were subjected to relatively similar soil and climatic conditions, and landscape characteristics, data generated within the catchment was used to assess the importance of different inherent properties and used amounts in quantifying pesticide occurrence in stream water (Paper IV). Stepwise multiple linear regression analysis was used to express concentration, transported amount and loss rate as functions of different intrinsic pesticide properties (or a combination of these) and quantities applied. The results demonstrated that the quantity applied in the catchment area was the most important estimator of pesticide concentrations and transported amounts. Model performance was slightly improved by adding intrinsic properties of the pesticides. Log Pow was the most significant intrinsic property for estimating relative loss of pesticides from the catchment. Using the best model equations in a cross-validation procedure, surface water concentrations, transported amount and loss rates were calculated that compared well to monitoring results for 80-100% of the compounds measured.

The conclusion that there was a strong association between pesticide occurrence in water and rates of pesticide use was in accordance with other studies. For example, Kolpin et al. (1998) found statistically significant relations between observed frequencies of pesticide detections in shallow groundwater and the use of these compounds. By using multiple regression analysis they concluded that, among the parameters considered, estimated use and Koc had the most profound effect on determining pesticide occurrence in groundwater. However, they also concluded that if site specific information on chemical use had been available, rather than the estimated use throughout the surrounding counties, the relation between chemical use and pesticide detection frequencies in groundwater might have been even stronger.

Battaglin & Goolsby (1998) applied regression equations to estimate herbicide concentrations in streamflow from large drainage basins as a function of different climatic conditions, soil type, topography, and land and herbicide use. The dominant variable, of those investigated, controlling herbicide concentrations in streamflow was found to be the rates of herbicide use within the drainage basins. They concluded that recommendations to reduce the amounts applied and to implement methods for minimising their loss to surface and groundwater systems would very likely lower pesticide concentrations in surrounding waters.

In a comprehensive review of groundwater monitoring studies by Barbash & Resek (1996), the authors concluded that literature values for intrinsic pesticide properties, such as soil dissipation half-lives and Koc, provided unreliable predictions of the pesticides that were most likely to be detected, or not detected in groundwater. The distribution of detected and non-detected pesticides with different dissipation half-lives and Koc values were found to be essentially random.

Larsson (1999) explored the influence of Koc and DT50 for a range of different pesticides on leaching behaviour in a structured clay soil in a modelling exercise. He demonstrated that for pesticides that were leached at rates of between 0.0001 and 10% of the applied dose, the influence of pesticide properties was significantly reduced due to macropore flow.

The well-structured, sandy loam soils in the Vemmenhög catchment exhibit a high infiltration capacity, which promotes rapid transport to drainage tiles. The findings in Paper IV support the concept of a strong relationship between amounts used and pesticide occurrence in stream water, as well as a minor influence of pesticide intrinsic properties, especially if derived from short-term laboratory tests or field studies representing other soil and climatic conditions.

Table 2. A rainfall loading estimate of the herbicide MCPA in relation to application load and streamflow transport based on data generated in 1991 (Papers I and II)
Process Total amount
(g)
Loada
(g/m2)
Fractionb
(%)
Spray application 188000 21000 -
Streamflow transport 223 25 0.1
Rainfall deposition 23 2.6 0.01

a Calculated for the entire catchment.

b Fraction of the applied volume in the catchment.

Importance of atmospheric deposition for measured concentrations in surface waters

Assuming the concentrations in the atmospheric deposition to be evenly spread, the deposition data presented in Paper I offered an opportunity to estimate approximate amounts of atmospheric pesticide deposition over larger areas and regions. For example, the total deposition of the investigated pesticides over Sweden's most southern province Scania (total area ∼ 11,000 km2) during the summer was calculated to correspond to 30-225 kg/season. This is equal to 26-199 mg/ha, which constitutes a small fraction of what is normally applied to an agricultural field. For example, the average amount of pesticides applied in the Vemmenhög catchment during spring and early summer ranged between 0.86-2.36 kg/ha during 1990-1996, with an overall average of 1.48 kg/ha (corresponding to 1.13 kg/ha if averaged over the entire catchment). A rough estimation would demonstrate that the atmospheric deposition constituted ca. 0.01% of the normal application rate. However, far from all pesticides used in the catchment, or in Sweden, were included in the deposition load calculations Consequently, the pesticide load is probably underestimated.

The contribution of atmospheric deposition to the amount of pesticides found in surface waters was calculated based on data generated in 1991 for the herbicide MCPA within the rainwater and Vemmenhög projects (Table 2). The total amount of MCPA applied in the Vemmenhög catchment (Paper II) was 188 kg, which corresponds to an applied amount for the entire catchment of 21 000 g/m2. The total amount of MCPA transported in streamflow leaving the catchment during May-September was 223 g, i.e. 25 g/m2, which corresponded to ∼ 0.1% of the applied amount (Paper II). The total deposition of MCPA in rainfall over the entire catchment was 23 g, based on a deposition load of 2.56 g/m2 measured at Ekeröd (Paper I). This corresponded to ∼ 0.01% of the applied amount in the catchment and, if this was carried directly to the stream, ∼ 10% of the streamflow load.

One of the pesticides most frequently detected in rainfall was the phenoxy acid herbicide 2,4-D, which was withdrawn from the Swedish market in 1989. Use of old stock was allowed after that date and in 1990 a total amount of 3 kg 2,4-D was applied in the Vemmenhög catchment; none was applied in 1991. The total amount of 2,4-D transported in streamflow was 7 g in 1990, corresponding to ∼ 0.2% of the applied amount, and there were no traces in streamflow during 1991. The total deposition of 2,4-D in rainfall at Ekeröd was 0.83 and 0.74 g/m2 in 1990 and 1991, respectively, corresponding to a total deposition over the catchment of 7.5 and 6.7 g in 1990 and 1991, respectively. This was equal to the amount found in streamflow (7 g) in 1990.

Hatfield et al. (1996) compared annual rainfall loadings of the herbicide atrazine with applied amounts within a catchment located in the mid-western US and found that rainfall loadings corresponded to < 0.1% of amounts applied. In another study, Goolsby et al. (1997) estimated the amounts of atrazine and alachlor deposited over the mid-western and north-eastern US to represent 0.6% and 0.4%, respectively, of the amount applied annually in the study area. These estimates are therefore somewhat higher than those based on the studies presented in Papers I and II.

Atmospheric deposition of pesticides to large surface water bodies can be substantial and its relative importance compared to other non-point sources is, in general, proportional to the surface area of the water body in relation to its terrestrial drainage area. Therefore a lake with a large surface area with respect to its drainage area usually receives much of its total inflow of water from direct wet and dry deposition and is therefore vulnerable to atmospheric contaminants (Eisenreich et al., 1981). Pacyna et al. (1993) estimated the total annual load for a number of organic chemicals to the Great Lakes as a result of atmospheric deposition as follows: atrazine ∼ 34 tons, trifluralin ∼ 3 tons, &aplha;-HCH ∼ 1 tons, lindane (γ-HCH) ∼ 0.5 tons, Σ endosulfan ∼ 0.3 tons and Σ DDT ∼ 0.1 tons.

Significance for the environment

The mere presence of a pesticide in the aquatic environment does not in itself constitute a risk to the flora and fauna living there. In order to make a proper risk assessment it is important to know both the exposure and the toxicity of the compound to relevant organisms in the environment. However, there has been a well recognised lack of adequate toxicity studies carried out under natural environmental conditions. Also, water-quality criteria for the protection of aquatic life, associated with the presence of current generation pesticides, have not been established either within Sweden or on a EU-level.

Table 3. Maximum Permissible Concentrations and Negligible Concentrations (Crommentuijn et al., 1997) for pesticides detected in the Vemmenhög Stream (Paper II)
Substance MPCa
(g/l)
NCb
(g/l)
Max konc.c
(g/l)
No weeks
> MPC
Atrazine 2.9 0.03 3 1
Bentazone 64 0.6 5 0
Chloridazon 73 0.7 20 0
Cyanazine 0.19 0.002 10 26
2,4-D 9.9 0.1 10 1
Deltamethrin 0.0003 0.000003 0d -
Dichlorprop 40 0.4 25 0
Dimethoate 23 0.2 30 1
Diuron 0.43 0.004 0.6 1
Isoproturon 0.32 0.003 10 37
MCPA 1.7 0.02 39 15
Mecoprop 3.9 0.04 16 9
Methabenzthiazuron 1.8 0.02 30 6
Metamitron 10 0.1 60 6
Metazachlor 34 0.3 200 3
Pirimicarb 0.09 0.0009 7 45
Simazine 0.14 0.001 10 29

a = Maximum Permissible Concentration

b = Negligible Concentration

c = Maximum weekly average concentration measured at site UT10 1992-1997 (during 122 weeks)

d = Limit of determination was 0.1 g/1

Dutch water-quality guidelines for the protection of aquatic life were recently suggested for 70 different pesticides (Crommentuijn et al., 1997). In this report, values for Maximum Permissible Concentration (MPC) and Negligible Concentration (NC) were given for 17 of the 37 pesticides detected in stream water from the Vemmenhög catchment (Table 3). The NC-value was derived by dividing the MPC-value by a factor of 100 to account for combination toxicity. In Table 3, MPCs and NCs are compared with concentrations measured in the Vemmenhög catchment; concentrations of five of the pesticides exceeded MCP during 15 to 45 weeks, i.e. 12-37% of the period under investigation.

The most clearly documented effects of atmospheric pesticides on human health and aquatic life are related to long-lived, environmentally stable organochlorine insecticides that concentrate in organisms through biomagnification (food chain accumulation), bioconcentration (partitioning), or both (Majewski & Capel, 1995). These persistent organic pollutants (POPs) arc transported world-wide and are frequently found at low levels in the aquatic environment, and can concentrate to significant levels in animals and humans. Determining the environmental significance of the concentrations of non-organochlorine pesticides in current use observed in atmospheric deposition is difficult due to a lack of adequate effect studies and guidelines for these matrices.

Concluding remarks and future research needs

Results from precipitation measurements show that there is significant contribution from modem agricultural pesticides via atmospheric deposition. Part of the deposited load is due to long-range transboundary deposition, originating from other parts of Europe. However, up to now the study presented in Paper I, which was carried out at three sites, contains the only reported measurements of pesticides in current use found in Swedish precipitation. This presents uncertainties in calculations of the actual deposition loads. Consequently, it is essential to collect more data from other parts of Sweden and also to extend the sampling periods. This would give an additional opportunity to study long-range transport of pesticides, since the application season starts much earlier, and ends later in countries further south.

The sampling strategy selected in the catchment study, where sampling intervals were time-paced and not flow-proportional, might have underestimated pesticide losses in streamflow. However, time-paced sampling was preferable from an environmental protection point of view, since aquatic life is vulnerable to pesticide concentrations occurring in the stream over time, and not to the amount of water passing the habitat. Furthermore, it could not be assumed that peak concentrations would only occur in connection with changes in streamflow, since other possible sources investigated were pesticide residues entering the stream as a result of wind-drift and accidental spills, which are not associated with changes in streamflow. Another consideration regarding the choice of sampling strategy was the aim of covering the whole growing season, which was more easily achieved if the available resources were distributed evenly over the season. Flow-proportional sampling requires a greater flexibility. Nevertheless, it is important to collect flow-proportional data to better understand the flow-processes and improve the calculations of actual pesticide losses in streamflow.

In order to obtain accurate load estimates, either in atmospheric deposition or streamflow, it is crucial that the analyses are carried out at a low level of detection. Low levels of pesticides occurring in large volumes of water can significantly contribute to the total load, such as during the winter. However, pesticide concentrations diluted in large volumes of water can fall below the limit of detected and thus become undetected. Also, as mentioned above, Dutch waterquality guidelines for the protection of aquatic life reveals that for several of the pesticides studied their respective level of detection exceeds, or is very close to the level at which toxic effects of sensitive aquatic organisms may occur. This means that for some of the investigated pesticides no definite statement can be made of their potential impact on aquatic flora and fauna until lower levels of detection have been achieved. Therefore, to do accurate load estimates and proper risk assessment, the methods of analysis need to be more sensitive, which requires improved analytical techniques.

Twenty years ago, there was a general lack of knowledge about the occurrence of modem pesticides in the environment at a distance from the point of application. This lack of information could partly be ascribed to the lack of appropriate analytical methods for this heterogeneous and complex group of organic pollutants. Today, analytical methodology has improved and residues of a range of current generation pesticides have been detected world-wide in a variety of different environmental compartments far away from emission sources. However, analyses are often focused on a few widely used compounds, e.g. the herbicide atrazine. Other pesticides are far less investigated due to large analytical difficulties. The lack of simple and inexpensive analytical methods is still a major obstacle doing adequate exposure and toxicity studies, and, consequently, also for risk assessment.

A lot of information can be obtained by studies in the laboratory and in the field prior to official approval and widespread use of a pesticide. Nevertheless, there must also be an adequate monitoring programme to identify environmental effects and occurrence in drinking water after the registration process is completed to preclude possible, unwanted environmental behaviour. In view of the normal seasonal variations in concentration, combined with year-to-year variations caused by differences in weather and agricultural practices, it is important to carry out long-term studies with a consistent methodology.

Simply basing the assessments of the risk of pesticide occurrence in water sources upon the amounts applied and the intrinsic properties of a pesticide can only give a very generalised indication. The agricultural system is heterogeneous and the fate of pesticides is therefore determined by the interaction of a range of complex factors often difficult to measure. However, to prevent surface water and groundwater contamination it is important to enhance our understanding of the different processes controlling pesticide movement in the environment.

References

Åkerblom, M. 1991. Pesticides in water - past and future national monitoring programs in Sweden. In (ed. A. Helweg) Pesticides in the aquatic environment - Appearance and effect. No. S 2181-1992,69-74. Tidskrift for Planteavis Specialserie, Denmark.

Albanis, T.A., Danis, T.G. & Kourgia, M.K. 1994. Transportation of pesticides in estuaries of the Axios, Loudias and Aliakmon rivers (Thermaikos Gulf), Greece. Sci. Total Environ. 156,11-22.

Arvidsson, T. 1997. Spray drift as influenced by meteorological and technical factors. Thesis Agraria 71. 144 pp. Swedish University of Agricultural Sciences, Department of Crop Production Science, Uppsala.

Atkinson, R., Kwok, E.S.C. & Arey, J. 1992. Photochemical processes affecting the fate of pesticides in the atmosphere. In Brighton Crop Protection Conference- Pests and Diseases, Vol. 2, 469-476. BCPC, Farnham, Surrey, UK.

Barbash, J.E. & Resek, E.A. 1996. Pesticides in Ground Water. Distribution, Trends, and Governing Factors. In (series ed. R.J. Gilliom) Pesticides in the Hydrologic System, Vol. 2. Ann Arbor Press, Chelsea, MI, USA. 588 pp.

Barceló, D., Chiron, S., Fernandez-Alba, A., Valverde, A. & Alpendurada. M.F. 1996. Monitoring pesticides and metabolites in surface water and groundwater in Spain. In (eds M.T. Meyer & E.M. Thurman) Herbicide Metabolites in Surface Water and Groundwater, ACS Symposium Series 630, 237-253. American Chemical Society, Washington, DC.

Battaglin, W.A. & Goolsby, D.A. 1998. Regression models of herbicide concentrations in outflow from reservoirs in the mid-western USA, 1992-1993. J. Am. Water Resour. Assoc. 34, 1369-1390.

Battaglin, W.A., Goolsby, D.A. & Coupe, R.H. 1993. Annual use and transport of agricultural chemicals in the Mississippi River, 1991-92. Open-File Report 93-418, 26-40. U.S. Geological Survey, Denver, CO, USA.

Bergström, L. & Stenström, J. 1998. Environmental fate of chemicals in soil. Ambio 27, 16-23.

Bester, K. & Hühnerfuss, H. 1993. Triazines in the Baltic and North Sea. Mar. Pollut. Bull. 26, 423-427.

Bidleman, T.F. & McConnell, L.L. 1995. A review of field experiments to determine air-water gas exchange of persistent organic pollutants. Sci. Total Environ. 159, 101-117.

Brown, C.D., Carter, A.D. & Hollis, J.M. 1995a. Soils and pesticide mobility. In (eds T.R. Roberts & P.C. Kearney) Environmental Behaviour of Agrochemicals, 9, 131-184. John Wiley & Sons, Chichester, UK.

Brown, C.D., Hodgkinson, R.A., Rose, D.A.. Syers, J.K. & Wilcockson, S.J. 1995b. Movement of pesticides to surface waters from a heavy clay soil. Pestic. Sci. 43, 131-140.

Buhler, D.D., Randall, G.W., Koskinen, W.C. & Wyse, D.L. 1993. Atrazine and alachlor losses from subsurface tile drainage of a clay loam soil. J. Environ. Qual. 22, 583-588.

Burgoa, B. & Wauchope, R.D. 1995. Pesticides in run-off and surface waters. In (eds T.R. Roberts & P.C. Kearney) Environmental Behaviour of Agrochemicals, 9, 221-255. John Wiley & Sons, Chichester, UK.

Crommentuijn, T., Kalf, D.F., Polder, M.D., Posthumus, R. & van de Plassche, E.J. 1997. Maximum permissible concentrations and negligible concentrations for pesticides. Report 601501 002. 174 pp. National Institute of Public Health and the Environment (RTVM), Bilthoven, The Netherlands.

Dekker, L.W., Ritsema, C.J., Wendroth, O., Jarvis, N., Oostindie, K., Pohl, W., Larsson, M. & Gaudet, J.P. 1999. Moisture distributions and wetting rates of soils at experimental fields in the Netherlands, France, Sweden and Germany. J. Hydrol. 275,4-22.

Dick, E.M. 1994. Water and wastewater sampling for environmental analysis. In (ed. B. Markert) Environmental Sampling for Trace Analysis, 255-278. VCH, Weinheim, Germany.

Domagalski, J. 1997. Results of a prototype surface water network design for pesticides developed for the San Joaquin River Basin, California. J. Hydrol. 192,33-50.

Eisenreich, S.J. 1990. Wet and dry deposition of semi volatile organic chemicals. In Workshop on long-range transport of organic pollutants, Report No 3807, 83-86. Swedish Environmental Protecion Agency, Stockholm, Sweden.

Eisenreich, S.J., Baker, J.E., Franz, T., Swanson, M., Rapaport, R.A., Strachan, W.M.J. & Hites, R.A. 1992. Atmospheric deposition of hydrophobic organic contaminants to the Laurentian Great Lakes. In (ed. J.L. Schnoor) Fate of Pesticides and Chemicals in the Environment, 51-78. John Wiley & Sons, Chichester, UK.

Eisenreich, S.J., Looney, B.B. & Thornton, J.D. 1981. Airborne organic contaminants in the Great Lakes ecosystem. Environ. Sci. Technol. 15, 30-38.

Eke, K.R. 1996. Pesticides in the aquatic environment in England and Wales. Pesticide Outlook April, 15-20.

Elliott, J.G. & Wilson, B.J. (eds) 1983. The Influence of Weather on the Efficiency and Safety of Pesticide Application. The Drift of Herbicides. Vol. 3. 135 pp. British Crop Protection Council, Thornton Heath, Surrey, UK.

Flury, M. 1996. Experimental evidence of transport of pesticides through field soils - A review. J. Environ. Qual. 25,25-45.

Frank, R., Braun, H.E., van Hove Holdrinet, M., Sirons, G.J. & Ripley, B.D. 1982. 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. 11, 497-505.

Frank, R., Clegg, B.S. & Patni, N.K. 1991a. Dissipation of atrazine on a clay'loam soil, Ontario, Canada, 1986-90. Arch. Environ. Contain. Toxicol. 21,41-50.

Frank, R., Clegg, B.S. & Patni, N.K. 1991b. Dissipation of cyanazine and metolachlor on a clay loam soil, Ontario, Canada, 1987-1990. Arch. Environ. Contam. Toxicol. 21, 253-262.

Frank, R., Logan, L. & Clegg, B.S. 1991c. Pesticide and polychlorinated biphenyl residues in waters at the mouth of the Grand, Saugeen, and Thames Rivers, Ontario, Canada, 1986-1990. Arch. Environ. Contam. Toxicol. 21,585-595.

Garmouma, M., Teil, M.J., Blanchard, M. & Chevreuil, M. 1998. Spatial and temporal variations of herbicide (triazines and phenylureas) concentrations in the catchment basin of the Mame river. Sci. Total Environ. 224,93-107.

Gish, T.J., Shinnohammadi, A., Helling, C.S., Kung, K.-J.S., Wienhold, B.J. & Rawls, W.J. 1998. Mechanisms of herbicide leaching and volatilization and innovative approaches for sampling, prediction and control. In (eds J.L. Hatfield, D.D. Buhler & B.A. Stewart) Integrated Weed and Soil Management, 107-134. Ann Arbor Press, Chelsea, MI. USA.

Gomme, J.W., Shurvell, S., Hennings, S.M. & dark, L. 1991. Hydrology of pesticides in a chalk catchment: surface waters. J. Inst. Water Environ. Manag. 5,546-552.

Goolsby, D.A., Thurman, E.M., Pomes, M.L., Meyer, M.T. & Battaglin, W.A. 1997. Herbicides and their metabolites in rainfall: Origin, transport, and deposition patterns across the midwestem and northeasten United States, 1990-1991. Environ. Sci. Technol. 31,1325-1333.

Griffini, O., Bao, M.L., Barbieri, C., Burrini, D. & Pantani, F. 1997. Occurrence of pesticides in the Amo River and in potable water - A survey of the period 1992-1995. Bull. Environ. Contam. Toxicol. 59,202-209.

Hagenvall, H. 1990. Droppar på drift. Vindavdrift vid besprutning. Aktuellt från lantbruksuniversitet 389. 31 pp. Sveriges lantbruksuniversitet, Uppsala. (In Swedish.)

Hallberg, G.R. 1989. Pesticide pollution of groundwater in the humid United States. Agric. Ecosystems Environ. 26, 299-367.

Harris, G.L., Jones, R.L., Catt, J.A., Mason, D.J. & Arnold, D.J. 1995. Influence of agricultural management and pesticide sorption on losses to surface waters. In (eds A. Walker, R. Allen, S.W. Bailey, A.M. Blair, C.D. Brown, P. Günther, C.R. Leake & P.H. Nicholls) Pesticide Movement to Water, BCPC Monograph No 62, 305-310. British Crop Protection Council, Farnham, UK.

Harris, G.L., Nicholls, P.H., Bailey, S.W., Howse, K.R. & Mason, D.J. 1994. Factors influencing the loss of pesticides in drainage from a cracking clay soil. J. Hydrol. 159, 235-253.

Hatfield, J.L., Wesley, C.K., Prueger, J.H. & Pfeiffer, R.L. 1996. Herbicide and nitrate distribution in central Iowa rainfall. J. Environ. Qual. 25, 259-264.

Hessel, K., Kreuger, J. & Ulén, B. 1997. Kartläggning av bekämpningsmedelsrester i ytgrund- och regnvatten i Sverige 1985-1995. Resultat från monitoring och riktad provtagning. Ekohydrologi 42,1-74. (In Swedish.)

Jarvis, N. 1998. Modeling the impact of preferential flow on nonpoint source pollution. In (eds H.M. Selim & L. Ma) Physical Nonequilibrium in Soils. Modeling and Application, 195-221. Ann Arbor Press, Chelsea, MI, USA.

Jaynes, D.B., Hatfield, J.L. & Meek, D.W. 1999. Water quality in Walnut Creek watershed: Herbicides and nitrate in surface waters. J. Environ. Qual. 28,45-59.

Jaynes, D.B., Hatfield, J.L. & Soenksen, P.J. 1994. Water and chemical transport in surface and tile discharge of Walnut Creek. In Toxic Substances and the Hydrologic Sciences, 430-445. American Institute of Hydrology, USA.

Johnson, A.C., Batchelor, C.H. & Haria, A.H. 1995a. Processes influencing pesticide transport in a drained clay catchment. In Effects of Scale on Interpretation and Management of Sediment and Water Quality, IAHS Publ. no 226, 39-46.

Johnson, D.C., Selim, H.M., Ma, L., Southwick, L.M. & Willis, G.H. 1995b. Movement of atrazine and nitrate in Sharkey soil: Evidence of preferential flow. Report No: 846. 23 pp. Louisiana State University Agricultural Center, Baton Rouge, LA, USA.

KemI 1998. Försålda kvantiteter av bekämpningsmedel 1997. 40 pp. National Chemicals Inspectorate, Solna. (In Swedish, with English summary.)

Kladivko, E.J., Van Scoyoc, G.E., Monke, E.J., Oates, K.M. & Pask, W. 1991. Pesticide and nutrient movement into subsurface tile drains on a silt loam soil in Indiana. J. Environ. Qual. 20, 264-270.

Klassen, W. 1995. World food security up to 2010 and the global pesticide situation. In (eds N.N. Ragsdale, P.C. Kearney & J.R. Plimmer) Eighth International Congress of Pesticide Chemistry. Options 2000, 1-32. American Chemical Society, Washington, DC.

Kolpin, D.W., Barbash, J.E. & Gilliom, R.J. 1998. Occurrence of pesticides in shallow groundwater of the United States: Initial results from the national water-quality assessment program. Environ. Sci. Technol. 32, 558-566.

Kreuger, J. & Brink, N. 1988. Losses of pesticides from arable land. Växtskyddsrapporter Jordbruk 49,50-61.

Larson, S.J., Capel, P.D., Goolsby, D.A., Zaugg, S.D. & Sandstrom, M.W. 1995. Relations between pesticide use and riverine flux in the Mississippi River basin. Chemosphere 31, 3305-3321.

Larson, S.J., Capel, P.D. & Majewski, M.S. 1997. Pesticides in Surface Waters. Distribution, Trends, and Governing Factors. In (series ed. R.J. Gilliom) Pesticides in the Hydrologic System, Vol. 3. Ann Arbor Press, Chelsea, MI, USA. 373 pp.

Larsson, M. 1999. Quantifying Macropore Flow Effects on Nitrate and Pesticide Leaching in a Structured Clay Soil. Field experiments and modelling with the MACRO and SOILN models. Thesis Agraria 164. Swedish University of Agricultural Sciences, Department of Soil Sciences, Uppsala.

Lee, D.S. & Nicholson, K.W. 1994. The measurement of atmospheric concentrations and deposition of semi-volatile organic compounds. Environmental Monitoring and Assessment 32,59-91.

Leonard, R.A. 1990. Movement of pesticides into surface waters. In (ed. H.H. Cheng) Pesticides in the Soil Environment: Processes, Impacts, and Modeling, Number 2, 303-349. Soil Science Society of America, Madison, Wisconsin, USA.

Lundbergh, I., Kreuger, J. & Johnson, A. 1995. 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. 54 pp. Council of Europe Press, ISBN 92-871-2776-X.

Majewski, M.S. 1991. Sources, movement and fate of airborne pesticides. In (ed. H. Frehse) Pesticide Chemistry. Advances in International Research, Development and Legislation. Seventh International Congress of Pesticide Chemistry, 307-317. VCH, Weinhem, Germany.

Majewski, M.S. & Capel, P.D. 1995. Pesticides in the Atmosphere. Distribution, Trends, and Governing Factors. In (series ed. R.J. Gilliom) Pesticides in the Hydrologic System, Vol. 1. Ann Arbor Press, Chelsea, MI, USA. 214 pp.

Mogensen, B.B. & Spliid, N.H. 1995. Pesticides in Danish watercourses: Occurrence and effects. Chemosphere 31, 3977-3990.

Muir, D.C. & Baker, B.E. 1976. Detection of traizine herbicides and their degradation products in tile-drain water from fields under intensive corn (maize) production. J. Agric. Food Chem. 24, 122-125.

Ng, H.Y.F. & Clegg, S.B. 1997. Atrazine and metolachlor losses in runoff events from an agricultural watershed: the importance of runoff components. Sci. Total Environ. 193, 215-228.

Pacyna, J.M., Bidleman, T., Evans, G., Keeler, G.J. & Voldner, E. 1993. Emissions, atmospheric transport, and deposition of heavy metals and persistent organic pollutants. Draft working paper for the first workshop on emissions and modeling of atmospheric transport of persistent organic pollutants and heavy metals. 89 pp. US Environmental Protection Agency, Durham, NC, USA.

Pereira, W.E. & Rostad, C.E. 1990. Occurrence, distributions, and transport of herbicides and their degradation products in the lower Mississippi River and its tributaries. Environ. Sci. Technol. 24,1400-1406.

Plimmer, J.R. 1988. Movement of pesticides from the site of application. In Pesticides: Food and Environmental Implications, IAEA-SM-297/35, 61-77. International Atomic Energy Agency, Vienna, Austria.

Rao, P.S.C., Jessup, R.E. & Davidson, J.M. 1988. Mass flow and dispersion. In (ed. R. Grover) Environmental Chemistry of Herbicides, J, 21-43. CRC Press, Boca Raton, FL, USA.

Richards, R.P. 1993. Composite sampling for toxics in drinking water. Great Lakes Protection Fund Project. Final Report FG6901064. 43 pp. Water Quality Laboratory, Heidelberg College, Tiffin, Ohio.

Richards, R.P. & Baker, D.B. 1993. Pesticide concentration patterns in agricultural drainage networks in the Lake Erie basin. Environ. Toxicol. Chem. 72,13-26.

Roberts, T.R. 1996. Assessing the environmental fate of agrochemicals. J. Environ. Sci. Health B 31, 325-335.

Rosling, D., Erlandsson, B., Pihlström, T. & Ericsson, B.-G. 1998. Dricksvattnet - stor undersökning av bekämpningsmedel. Vår Föda 1,22-28. (In Swedish.)

SCB 1997. Bekampningsmedel i jordbruket 1995/96. Förbrukning i åkergrödor. Na 31 SM 9702. 25 pp. Statistics Sweden, Stockholm. (In Swedish, with English summary.)

Schottler, S.P., Eisenreich, SJ. & Capel, P.D. 1994. Atrazine, alachlor, and cyanazine in a large agricultural river system. Environ. Sci. Technol. 28,1079-1089.

Schroeder, W.H. & Lane, D.A. 1988. The fate of toxic airborne pollutants. Environ. Sci. Technol. 22,240-246.

Seiber, J.N. & Woodrow, J.E. 1995. Origin and fate of pesticides in air. In (eds N.N. Ragsdale, P.C. Kearney & J.R. Plimmer) Eighth International Congress of Pesticide Chemistry. Options 2000, 157-172. American Chemical Society, Washington, DC.

SJV 1996. Förslag till handlingsprogram för att minska hälso- och miljöriskerna vid användning av bekämpningsmedel till 2001. Rapport 1996:2. 171 pp. Swedish Board of Agriculture, Jönköpning. (In Swedish.)

Skark, C. & Zullei-Seibert, N. 1995. The occurrence of pesticides in groundwater - Results of case-studies. Intern. J. Environ. Anal. Chem. 58, 387-396.

Spliid, N.H. & Køppen, B. 1998. Occurrence of pesticides in Danish shallow ground water. Chemosphere 37,1307-1316.

Strachan, W.M.J. & Huneault, H. 1984. Automated rain sampler for trace organic substances. Environ. Sci. Technol. 18,127-130.

Sunito, L.R., Shiu, W.Y., Mackay, D., Seiber, J.N. & Glotfelty, D. 1988. Critical review of Henry's law constants for pesticides. Rev. Environ. Contain. Toxicol. 103,1-59.

Taylor, A.W. 1995. The volatilization of pesticide residues. In (eds T.R. Roberts & P.C. Kearney) Environmental Behaviour of Agrochemicals, 9, 257-306. John Wiley & Sons, Chichester, UK.

Traub-Eberhard, U., Kördel, W. & Klein, W. 1994. Pesticide movement into subsurface drains on a loamy silt soil. Chemosphere 28,273-284.

Van Pul, W.A.J., de Leeuw, F.A.A.M., van Jaarsveld, J.A., van der Gaag, M.A. & Sliggers, C.J. 1998. The potential for long-range transboundary atmospheric transport. Chemosphere 37,113-141.

Wauchope, R.D. 1978. The pesticide content of surface water draining from agricultural fields - A review. J. Environ. Qual. 7, 459-472.

Wauchope, R.D., Graney, R.L., Cryer, S., Eadsforth, C., Klein, A.W. & Racke, K.D. 1995. Pesticide runoff: Methods and interpretation of field studies. Pure & Appl. Chem. 67, 2089-2108.

Wauchope, R.D. & Leonard, R.A. 1980. Maximum pesticide concentrations in agricultural runoff: A semiempirical prediction formula. J. Environ. Qual. 9,665-672.

Acknowledgements

Throughout the many years during which I worked to complete this thesis, I have had the pleasure to enjoy the company of numerous people who have helped and inspired me to accomplish this thesis. I want to express my sincere gratitude to you all. There are a also few I would like to mention specially:

First of all I would like to thank my three supervisors:

I am also indebted to:

Financial support for the work included in this thesis was provided by the Swedish National Chemicals Inspectorate, the Swedish Environmental Protection Agency, the Nordic Council of Ministers, the Environmental Foundation of the Malmöhus county council and the County Government Board in Malmöhus county, and is gratefully acknowledged.