| Contents |

Series no. XX, 2000

Report of the sub-committee on the environment and health.

Subtitle

Contents

Preface

The Sub-committee for Environment and Health, which is part of the Bichel Committee, was appointed in autumn 1997 to assess the consequences for the environment and public health of phasing out the use of pesticides. This report is a result of those assessments and is one of five technical background reports that form the basis for the Bichel Committee's final report to the Minister of Environment and Energy.

The other background reports cover the consequences for agriculture, the consequences for the economy and employment and, lastly, the legal possibilities of phasing out the use of pesticides. The last-mentioned report also covers the overall consequences of total restructuring for organic farming.

This is the first time in Denmark - and probably also internationally - that such an extensive interdisciplinary analysis has been conducted of the consequences for agriculture of a total or partial phase-out of pesticide use and of total restructuring for organic production.

The Sub-committee for Environment and Health found a wide range of health and environmental impacts from pesticides that are now banned. On this basis, the sub-committee also found that our knowledge about impacts is constantly increasing, so substances that are at present regarded as safe may be viewed differently in the future. Experience to date thus confirms that pesticides should only be used with great caution.

The sub-committee also ascertained a number of impacts from the use of pesticides on both the terrestrial and the aquatic environment, but did not find any health effects on the population in general from currently authorised pesticides. The sub-committee identified various lacunae in our knowledge that make it difficult to calculate the consequences of phasing out the use of pesticides. The sub-committee therefore had to limit the actual analysis of consequences to the few impacts about which there was sufficient background information.

The sub-committee also considered how the environment and health could be protected still further and, in that connection, indicated various areas for action.

The sub-committee based its report on a number of consultants' reports. The consultants, the members of the sub-committee and the secretariat all made a major contribution to the creation of the report, and we take this opportunity to thank everyone concerned for their good work.

Henrik Sandbech
Chairman of the Sub-committee for Environment and Health

1. Introduction

On 15 May 1997 the Folketing (the Danish Parliament) unanimously passed a parliamentary resolution urging the government to appoint a committee with independent expertise to analyse all the consequences of totally or partially phasing out the use of pesticides in agriculture and to examine alternative methods of preventing and controlling plant diseases, pests and weeds.

The committee was to assess the consequences for production, the economy, legislation, health and the environment, and employment. In continuation of its mandate (see chapter 2), the Sub-committee for Environment and Health looked at issues relating to the precautionary principle (see chapter 8) and the relationship between the pesticide load and the load from other chemical impacts in agriculture (see chapter 7).

The results of the committee’s work were to be used in the coming work on a new pesticide action plan.

In the mandate of 4 July 1998, the Minister of Environment and Energy stipulated that a main committee be appointed with expert members from the research world, the agricultural industries, the "green" organisations, consumer organisations, the foodstuffs and agrochemical industries, the trade unions and relevant ministries. Its members were to cover the specialist areas of agricultural production, economics, legislation, employment, health, the environment and ecology.

In addition, four sub-committees were appointed. Their task was to facilitate the main committee's final reporting by drafting specialist background reports.

The main committee had the task of coordinating and discussing the sub-committees' work and of preparing the final report for the Minister.

The sub-committees were to cover the following areas:

1. Agriculture
2. Production, economics and employment
3. Environment and health
4. Legislation

As points of reference for their work, the sub-committees were to use both the optimum production from the standpoint of operating economy and the production achieved to date by the agricultural industries – farming, market gardening and forestry. They were to assess the consequences for production, the economy, legislation, health, employment and the environment.

In their work, the sub-committees were to evaluate scenarios for total and partial phasing out of pesticides and examine the consequences of restructuring for organic farming, taking into account activities already in progress concerning such restructuring.

2. The sub-committee’s mandate and composition

2.1 The sub-committee’s mandate and scenarios for phasing out pesticides

According to its mandate, the Sub-committee on Environment and Health, working on the basis of the above-mentioned cultivation systems, was to evaluate the environmental consequences of a total or partial phase-out of pesticides. In its evaluation, the sub-committee was to consider the effects on groundwater as a resource for the population and the natural environment, surface water as a resource for flora and fauna, and the terrestrial ecosystems in agriculture and forestry as a resource for flora and fauna.

In its evaluation of the health consequences, the sub-committee was to include the effects of pesticides on the people using them and the effects of using the proposed cultivation systems. On the other hand, the sub-committee was not required to consider the health and environmental aspects of the industrial production of pesticides.

The sub-committee had the following members:

The President

Henrik Sandbech, Director General, National Environmental Research Institute (DMU)

Members

Professor Poul Bjerregård, Odense University
Dr. Hans Løkke, Director of Research Department, National Environmental Research Institute (DMU)
Dr. Agro. Arne Helweg, Danish Institute of Agricultural Sciences (DIAS)
Dr. Erik Thomsen, Government Geologist, Geological Survey of Denmark and Greenland
Dr. Ib Knudsen, Executive Director, Veterinary and Food Administration (VFA)
Dr. Ole Ladefoged, veterinarian, VFA
Dr. Lars Ovesen, Head of Division, VFA
Professor Finn Bro-Rasmussen, Technical University of Denmark (DTU)
Dr. Ib Johnsen, Institute of Botany, Organic Department, University of Copenhagen
Dr. Peter Jantzen, The National Board of Health (retired from the sub-committee in January 1998)
Elle Laursen, Deputy Superintendent, The National Board of Health (joined the sub-committee in January 1998)
Sonja Plough Jensen, Principal, Danish Working Environment Authority (WEA) (retired from the sub-committee in January 1998)
Kjeld Mann Nielsen, Special Consultant, WEA (joined the sub-committee in January 1998, retired in December 1998)
Bent Horn Andersen, Special Consultant, WEA (joined the sub-committee in December 1998)

In addition, Claus Vangsgård, M.Sc., Association of Danish Waterworks, participated in some of the sub-committee’s meetings.

The secretariat comprised Dr. Kaj Juhl Madsen, Danish Environmental Protection Agency, Lise Nistrup Jørgensen, Senior Scientist, Danish Institute of Agricultural Sciences, and Anne Marie Linderstrøm, Head of Section, Danish Environmental Protection Agency.

The sub-committee’s report was edited by Hans Løkke, Director of Research Department, National Environmental Research Institute, and Birgit Nygaard Sørensen, Executive Secretary, National Environmental Research Institute, took care of the layout and writing of the report.

3. Definition of the work

With a view to evaluating the consequences of a total or partial phase-out of pesticides, the sub-committee first gathered together the existing information on the known environmental and health effects of the use of pesticides as a reference frame for analysing the various cultivation systems set up by the Sub-committee on Agriculture.

Definition of pesticdes

In its work, the sub-committee used the definition of pesticides proposed by the Sub-Committee on Legislation. That definition is reproduced below:

"The expressions pesticides, plant protection products and active ingredients appear frequently in the following. For the sake of clarity, they and some other terms are therefore explained here.

Pesticides is an inclusive term for plant protection products and biocides.

Plant protection products (or spraying products) are the products used in the agricultural industries (farming, forestry, market gardening and fruit production) for prevention and control of weeds, pests and fungal diseases and for regulating growth (including straw shortening). Plant protection products can also – where permitted by law – be used in private gardens and similar, including uncultivated and public land.

Biocides include wood protection products, rat poisons and slimicides for use in paper pulp etc.

A pesticide consists of one or more active ingredients and some coformulants. The active ingredient is the substance that is specifically aimed at pests etc. The coformulants are substances that are added to enhance the intended effect."

The sub-committee knew that the Committee on Drinking Water appointed by the Danish Environmental Protection Agency in March 1996 had presented its report on 17 December 1997. The report showed that the committee had examined the existing regulation of pesticides and the risk of contamination of drinking water with pesticides. Therefore, where relevant, the Sub-committee on Environment and Health used this report as background material.

The sub-committee examined the existing body of knowledge to determine the general level of knowledge in this area and to ascertain how any lack of knowledge affected the possibility of evaluating the consequences of a total or partial phase-out of pesticides.

At the request of the main committee, the sub-committee also assessed the possibility of identifying the most harmful of the pesticides authorised for use today.

In that connection, the sub-committee considered how the precautionary principle could be operationalised in the case of pesticides and examined the pesticide problem in relation to other chemical compounds used in agriculture.

The sub-committee did not carry out a systematic evaluation of the present authorisation scheme, partly because this lay outside its mandate and partly because an expert evaluation had already been carried out in the case of groundwater protection in the spring and summer of 1997. Readers are also referred to a brief review in Annex 1 of the data required for authorisation of pesticides and to Annex 2 concerning the general rules for risk assessment of environmental impacts and classification of health impacts.

Consultants’ reports

Pesticides detected in the groundwater: Walter Brüsch, Bo Lindhardt, Martin Hansen, GEUS

Exposure and effect of pesticides on harmful insects in cultivated and terrestrial ecosystems: Lars Monrad Hansen, DIAS

Effects on flora in cultivated and uncultivated ecosystems: Gösta Kjellsson, DMU, Kathrine Hauge Madsen, KVL

Pesticides detected in precipitation, contribution of precipitation to soil and groundwater and international contribution to the pesticide load: Gitte Felding, DIAS

Pesticides detected in drain water and soil water: Niels Henrik Spliid, DIAS

Reduction potential at filling and washing yards: Niels Henrik Spliid, DIAS

Exposure and effects of pesticide usage on useful fauna and other lower and higher fauna in arable land: Niels Elmegård, DMU

Pesticides in soil: Inge S. Fomsgaard, DIAS

Effects on flora and fauna in fresh water: Nikolaj Friberg, DMU

Pesticides found in watercourses and lakes, together with surface run-off: Betty Bügel Mogensen, DMU

Ingestion of pesticides through diet: Arne Büchert, VFA

Effect of pesticides on public health: Lars Rauff Skadhauge, The National Board of Health

Analysis/description of the consequences for health and safety of phasing out pesticides within the agricultural industries: Karen Gertrud Bjørn, Occupational Health Service (OHS) Greater Copenhagen, Leif Rothmann, OHS Funen

Interaction between pesticides and production of toxins in food products: Susanne Elmholt, DIAS

Input concerning: Nitrate leaching during different forms of soil treatment: Elly Møller Hansen, Jørgen Djurhuus, DIAS

Change in consumption of fossil energy with restructuring for 0-pesticide agriculture: Tommy Dalgaard

Synthesis of the known effects of pesticides on the agro-ecosystem’s organisms and modelling of effects on bird populations on arable land in connection with a total or partial phase-out of pesticides: Bo Svenning Petersen, Flemming Pagh Jensen, Ornis Consult

Pesticide scenarios: Effects on lower fauna: Niels Elmegård, Jørgen A. Axelsen, DMU

Analysis of the environmental impact of pesticide use in private forestry: Morten Strandberg, DMU

Effect of pesticides on the aquatic environment with different treatment frequency indices: Flemming Møhlenberg, Kim Gustavson, DHI Water & Environment

Technical background report on identification of the most harmful pesticides authorised for use today: Bo Lindhardt, GEUS, Peter B. Sørensen, DMU, Niels Elmegård, DMU, Inge S. Fomsgård, DIAS, Else Nielsen, VFA, Jim Hart, VFA, Lene Gravesen, Danish Environmental Protection Agency (DEPA), Lerke Ambo Nielsen, DEPA, Lene Lorenzen, DEPA, Kaj Juhl Madsen, DEPA.

Comments received from individuals/institutions that were not members of the sub-committee

Novartis, Ole Jensen: Midway Conference in the Bichel Committee

Novartis/Mads Kristensen: Comments and questions to the Pesticide Committee

Bayer, Frank Rothweiler: Follow-up on the Midway Conference in the Bichel Committee

Cheminova, Lars-Erik K. Pedersen: Comments to the Committee following the Midway Conference in the Bichel Committee

Hans Sanderson, Comments to the Sub-committee for Environment and Health

Peter Højer: Questions to the Pesticide Committee

Jens Husby: Questions to the Bichel Committee

Institute for Product Development, Technical University of Denmark, Bente Mortensen

Comments to Henrik Sandbech’s paper at the Midway Conference in the Bichel Committee

Helga Moos, MP (Liberals): Comments to the Midway Conference

Zeneca Agro, Freddy Kjøng Pedersen: The Midway Conference on 21 September 1998

Erik Faurholt: Questions to the Pesticide Committee.

4. Occurrence of pesticides in the environment

Pesticides are distributed in the environment through spraying, spreading as granulate, with dressed seed, through swabbing, etc. During spraying, the pesticides can be carried by the wind over both short and long distances. Some chemicals can also evaporate from the surface of plants or soil. Atmospheric transport can be of considerable magnitude and is often far-reaching and transboundary. Pesticides can also be washed out from the atmosphere with rainwater or be deposited on surfaces by so-called dry deposition. Pesticides can spread with rainwater to the aquatic environment, where they occur in soil water, run-off water, drain water, watercourses and lakes. Lastly, pesticides can be conveyed to the environment by accident or through unlawful use.

This section describes the occurrence of pesticides in the different compartments, as shown in table 4.1. This table also shows the types of organism that are exposed to pesticides. Section 4.1.6 also describes how pesticides can be distributed in the environment and thus contribute to exposure of people and the environment. The possible effects of pesticides are described in section 4.2. The main effects occur in connection with the treatment with pesticides, when organisms are directly affected, but there are also indirect effects as a consequence of the impact on food chains. The report does not cover bioaccumulation of pesticides in organisms and possible concentration in food chains. The Danish Environmental Protection Agency’s assessment of the bioaccumulability of pesticides is certain to lead to such pesticides not being authorised in Denmark.

Table 4.1
Compartments in which pesticides occur and the types of organisms exposed to them. The table also shows the sections of the report in which the occurrence is described and the sections in which the potential effects of pesticides in the different compartments are described.

Systematic data are not available on the occurrence of pesticides in lakes and coastal waters, so the report does not include data from these compartments.

4.1 Pesticides in groundwater

Denmark’s water supply system is based on pure groundwater, which can be supplied to the consumers without treatment. In the last few decades, groundwater has been found to be contaminated with industrial chemicals, leachate from landfills, nitrate, heavy metals and, lately, pesticides. In 1994, contamination of the groundwater in Ejstrupholm with the herbicide atrazine led to splash headlines in the media. At the same time, countrywide analyses showed that pesticides were commonly found in the groundwater. New research results also show that even in clayey soil, fissures and cracks from the Ice Age enable pesticides to leach down to the groundwater relatively quickly.

Groundwater quality is one of the main elements of Denmark’s environmental legislation. For pesticides, Denmark uses the EU’s limit values of 0.1 microgramme per litre water for every substance and 0.5 microgramme per litre water for the sum of pesticides, irrespective of the nature of the substance. Since groundwater is regarded as a basic resource that must be kept free of pollution, these limit values have been set as "hygienic limit values". This is further supported by the fact that the groundwater used as drinking water forms over long periods of time. This means that by the time pollution can be measured in the aquifer itself, it is too late to avert it or control the extent of it.

Database

The groundwater monitoring programme (GRUMO) covers 67 monitoring areas spread around the country. The areas have been chosen with a view to providing a representative picture of Denmark’s aquifers. In most cases, the monitoring areas are situated in regions in which land is predominantly used for farming. Most of the areas include an extraction well to a waterworks and 10-20 specially established monitoring wells that characterise the main flow pattern of the groundwater (Figure 4.1).

The samples of water are extracted from well screens. A "well screen" should be understood to mean a closed pipe placed at the lower end of the groundwater well shaft. The well screen is equipped with slots or similar openings, which, in interaction with a screen pack of filter gravel around the well screen, are intended to retain the fine particles in the soil. The length of the well screen can vary so that the pumped up water comes from the depth interval in which the well screen is placed. Three types of well screen are used in the groundwater monitoring programme: point, line and volume monitoring screens. The last-mentioned type is often a waterworks well. In the point-monitoring and line-monitoring screens, the water samples represent the chemistry of the water samples at a specific point at a specific time in the aquifers examined. The volume-monitoring screens, on the other hand, represent mixed water, cf. the section on raw-water monitoring.

Figure 4.1 Look here!
Types of monitoring in the monitoring programme. GVS = groundwater table

The Nationwide Monitoring Programme for Agricultural Catchments (LOOP)

The Nationwide Monitoring Programme for Agricultural Catchments (LOOP) is carried out in five well-defined run-off catchment areas in agricultural regions where farming practice is known. Among other things, groundwater near the surface is monitored, so it is possible to relate leached substances to the use of the land. However, there are only limited data concerning pesticides from LOOP.

The water companies’ raw water monitoring system

The raw water monitoring system includes monitoring the water from the water companies’ extraction wells. In many cases, the interval in the wells from which water is extracted is of considerable length, and the water can be extracted from several separate saturated sediments. A water sample from the raw water monitoring system is therefore often a "mixed sample" of different types of water of different age and containing different substances. Since a number of water companies have extraction wells near towns, pesticides detected in the raw water monitoring system often bear signs of non-agricultural use.

In its yearly reporting of the results from the groundwater monitoring programme, GEUS gives the results from GRUMO, LOOP and the water companies’ raw water monitoring system separately because they concern different types of water sample.

Detection of a pesticide, or a metabolite, is defined as detection of the substance in question above the current detection limit. The laboratories use different detection limits. For example, DMU specifies a detection limit of 0.005 microgramme per litre for some substances, while other laboratories specify a detection limit of 0.01 microgramme per litre for the same substances. The detected substances are then divided into two groups: one over the detection limit and one over or equal to the limit value (i.e. the latter is a subset of the first group).

Pesticides have been found in groundwater all over the country with the exception of some areas north of the Limfjord and in North Zealand. Table 4.3 shows the pesticides detected in the raw water monitoring system.

The groundwater monitoring programme, GRUMO

The national groundwater monitoring programme gives a complete picture of the state of the groundwater, including the situation with respect to pesticides (GEUS 1998). Under the groundwater monitoring programme, analyses were carried out for eight pesticides in 1,014 well screens in the period 1990-1997. The programme covers two triazines (atrazine, simazine), four phenoxyacetic acids (dichlorprop, mechlorprop, MCPA, 2,4-D) and two nitrophenols (dinoseb, DNOC). Of the eight GRUMO pesticides, three (atrazine, dinoseb and DNOC) are now banned, while five of the others are subject to restrictions with respect to dosage, crops, etc. In all, 4,230 analyses of water samples have been carried out for these eight substances. As a consequence of the county authorities’ expanded analytical programmes, data were reported in 1998 from 594 well screens in the monitoring system from which water samples were analysed for more pesticides and their degradation products.

One or more of the eight GRUMO pesticides were detected one or more times in 121 well screens. The screens came from 101 wells, in 16 of which pesticides were found in two or three screens. The 121 screens correspond to just over 12% of the screens analysed, while the limit value for drinking water (0.1 microgramme per litre) was exceeded in 35 screens, corresponding to just under 3.5%, see table 4.2. There is often an interval of three years between sampling, and as pesticides normally occur in pulses, they are often not detected again in subsequent samples.

In the 594 screens in which some of the county authorities have carried out expanded analytical programmes, pesticides or degradation products have been detected in 21% and the limit value has been exceeded in 13%, see table 4.2. However, data have usually only been reported on one set of water samples from the 594 screens.

Counties analyse samples from sensitive wells for BAM

In the case of 2,6-dichlorobenzamide (BAM), which has been detected in about 14% of the screens analysed, the data must be treated with caution because the county authorities selected sensitive screens for testing for this compound. Sensitive screens should be understood to mean screens in young groundwater near the surface or screens close to potential sources of pollution. BAM is a degradation product, the parent compound of which is dichlobenil. Dichlobenil has not been used for agricultural purposes and is no longer permitted in Denmark.

The triazine degradation products desethylatrazine, desisopropyl-atrazine and hydroxyatrazine have been detected in 6.6%, 5% and 2.8%, respectively. However, hydroxyatrazine has been analysed for in only a few screens, and experience from other countries does not normally reveal high detection percentages for this substance.

Table 4.2
Detection of pesticides related to the number of analysed screens in the groundwater monitoring programme (GEUS 1998).

m g/L = microgramme per litre

The extension of the monitoring programme in 1998 to approx. 50 pesticides and degradation products may reveal possible further occurrence of pesticides in groundwater, particularly as only a few of the 594 screens have been analysed for all 50 substances.

The depth distribution of the pesticides shows that the eight GRUMO pesticides occur in 22% of the high-lying and youngest groundwater in the interval 0-10 metres below ground level and that the frequency decreases with the depth. Where screens have been analysed for more than eight substances in the monitoring programme, pesticides have been detected in 34% of the groundwater samples in the interval 0-10 metres below ground level, see figure 4.2. Of these, 23% of the detections were over or equal to 0.1 microgramme per litre.

It can thus be concluded that pesticides and their degradation products occur particularly in young groundwater. This may be due partly to the fact that the substances are gradually degraded while being transported down through the aquifers and partly to the fact that they are transported horizontally with the groundwater to watercourses, lakes and groundwater wells. It is also possible that the pesticides are transported towards the groundwater so slowly that the increased occurrence and rising concentrations of the substances found today in young ground water will be found in the lower aquifers in the future. It will only be possible to assess this when there are longer time series in the monitoring programmes, dating of groundwater with CFC and results from research projects that are being carried out to elucidate this.

Figure 4.2
Detections of pesticides in % against depth in metres below ground level are shown as white bars. The youngest groundwater is presumably most often found in the interval 0-10 metres below ground level. In analyses for eight GRUMO pesticides in the depth interval 0-10 metres below ground level, pesticides have been detected in 22% of the screens. The detection percentage rises to 34% when analysing for more than eight pesticides. In the bar chart, the black, horizontal bars show the detection percentage for the number of detections over or equal to 0.1 microgramme per litre (GEUS).

(Figure texts:
fundprocent = detection percentage
dybde interval i meter under terræn = depth in metres below ground level
pesticider = pesticides
dybde i meter under terræn = depth in metres below ground level
udvidet analyseprogram = expanded analytical programme

The water companies’ raw water monitoring system

In the water companies’ raw water monitoring system, pesticides have been detected in 17% of 4,209 wells, and the limit value has been exceeded in 6%. These wells have mainly been analysed for the eight GRUMO pesticides.

The most frequently analysed substances are shown in table 4.3. The number of analyses does not reflect the number of analysed wells; for example, dichlorprop has been analysed in 5,714 water samples from 3,998 wells. Of the eight GRUMO pesticides, dichlorprop, mechlorprop and atrazine have been detected most frequently, while, in percentage terms, degradation products from triazines have been detected more frequently in the water samples than, for example, atrazine. The relatively frequent occurrence of hexazinone is surprising because this substance has only been detected in samples from groundwater monitoring wells even though it has been used frequently in forestry. One explanation could be that hexazinone has been used for treating urban areas.

The distribution of finds of pesticides in the water companies’ wells is very reminiscent of the distribution in the groundwater monitoring system. However, the analysed pesticides occur in 26% of the water samples taken from wells with "top screen" in the interval from 10 to 20 metres below ground level, compared with 13% in the groundwater monitoring programme. This may be because the water companies in some areas extract groundwater from high-lying, fractured limestone, where the groundwater is presumably younger and more affected by the use of pesticides on the surface of the earth.

As a consequence of the many detections of the degradation product BAM (2.6-dichlorobenzamide) in the last few years, some water companies have analysed for this substance in their routine monitoring. The water companies have carried out 2,310 BAM analyses of water samples from 1,656 wells and have detected BAM in 448 of them, corresponding to around 30%. The limit value for drinking water was exceeded in 187 wells, corresponding to about 11% of the analysed wells. Most of the detections of BAM in waterworks wells were made from ground level to a depth of 0-30 metres, where the highest concentrations were also measured. BAM’s parent compound is a herbicide, dichlobenil, which has primarily been used in urban areas. Because of the earlier widespread use of the now prohibited parent compound dichlobenil, BAM can thus be expected to occur both under built-up areas and under farmyards, gravel roads and other areas kept free of vegetation.

Expanded analytical programmes

In recent years, the county authorities, the water companies and DMU have carried out a number of expanded analytical programmes for groundwater and waterworks water. Results covering 108 pesticides and metabolites are available. Of these 108 substances, around 40 have been detected in Danish groundwater, 29 of them in concentrations above the limit value for drinking water, see table 4.4. DMU's expanded analytical programmes comprise mainly analyses of groundwater from the LOOP regions. The data in the table are from analyses carried out in 1997, from older analyses that were not reported to the groundwater database at GEUS, and from analyses that were partly or totally reported to the groundwater database at GEUS, e.g. analytical programmes carried out in LOOP and a few groundwater monitoring areas. The table does not include data from the general monitoring of the groundwater monitoring areas. Twenty of the pesticides detected in groundwater were in use in Denmark in 1996, but some of them have since had restrictions imposed on their use or have been banned. However, a few substances, e.g. the degradation product ETU, was found in both soil water and groundwater in a research project by Fladerne Bæk (river) and in an analysis carried out by Copenhagen County Council in 1997, in which ETU was detected in piezometric wells close to a landfill site.

Table 4.3 Look here!
Detections of pesticides analysed in connection with the water companies’ raw water monitoring system (Brüsch et al. 1998). Ranked by falling number of analyses. Consumption data from Bekæmpelsesmiddelstatistik (Pesticide Statistics) 1997 (Danish Environmental Protection Agency 1998a).

Table 4.4 Look here!
Pesticides detected in groundwater in county authorities, water companies and DMU’s expanded analytical programmes. The situation in March 1998 from GEUS (Brüsch et al. 1998). Consumption data from Bekæmpelsesmiddelstatistik (Pesticide Statistics) 1997 (Danish Environmental Protection Agency 1998a).

In connection with a search of the literature for analytical data from foreign monitoring programmes for pesticides in groundwater, a search was made for 544 pesticides and metabolites that have been or still are used in Denmark. Of these substances, data were gathered on 281 that have been analysed abroad, 159 of which have been found mainly in groundwater. About 55 of these were used in Denmark in 1996. However, for many of these substances, there have been only a few finds in groundwater.

Finds of pesticides under different types of area

Finds of pesticides in deep aquifers cannot be definitely related to specific fields or other areas within the individual catchment area boundaries.

Prompted by a dialogue with the Agricultural Advisory Centre, GEUS has calculated the impact of pesticides on groundwater close to the surface of the ground in rural areas (including forests, wetlands, etc.) on the basis of data from the groundwater monitoring programme. The data treatment is conducted in relation to permitted and banned pesticides and in relation to agricultural and non-agricultural pesticides. Hydroxy-terbuthylazine was not included in the calculation owing to too little data. All data from the City of Copenhagen, Frederiksberg Municipality and Copenhagen County were omitted. Only screens placed in the interval 0-10 metres below ground level were included because it is estimated that groundwater near the surface is often the youngest groundwater. However, the age of this groundwater can vary.

Tables 4.5-4.7 show that pesticides found in the groundwater monitoring programme in rural areas in groundwater close to the surface are equally distributed between agricultural and non-agricultural substances. Most of the substances found are now banned or regulated with restrictions on their use.

Table 4.5
Permitted and banned pesticides. Of the 4 phenoxy acids, only MCPA is included as permitted on account of a large consumption, while the other 3 are included as banned substances even though they are still sold in small quantities. Desisopropylatrazine is included with 50% in each group because the metabolite can come from pesticides permitted today.

m g/L = microgramme per litre

Table 4.6
Agricultural/non-agricultural pesticides. All triazines are equally distributed between agricultural and non-agricultural because these substances have also been used in both forestry and fruit growing and in urban areas. However, simazine and hexazinone have not been included as agricultural substances. In this calculation, GEUS has not differentiated between permitted and banned substances.

m g/L = microgramme per litre

Table 4.7 Look here!
Background data, groundwater close to ground level. Groundwater monitoring data. Urban areas omitted. Only screens 0-10 metres below ground level are included. P = permitted; R = permitted with restricted use or authorised with conditions; B = banned; A = agricultural use; NA = non-agricultural use; Fr = fruit, berries, forest, plantation, etc. "Number of screens ³ 0.1 m g/L" is a sub-set of "number of screens with finds" (m g/L = microgramme per litre).

Table 4.8 Look here!
Pesticides and metabolites found in ground water in the five nationwide monitoring catchment areas during the period 1990-1997. The median value has been calculated on the basis of median values at screen level. GRUMO pesticides are in bold (GEUS 1998) (m g/L = microgramme per litre).

Triazines (including atrazine) and their metabolites are included in the calculation as equally distributed between use in agriculture and in urban areas. However, since agricultural land predominates in the groundwater monitoring programme, these substances should presumably be weighted higher as agricultural use. This is supported by the available data on pesticide finds in groundwater near ground level in the LOOP areas, where only agricultural land with known cultivation practice is monitored. There, high find percentages are found, particularly in the case of triazines’ metabolites (Table 4.8).

4.1.2 Conclusions

In Denmark, extensive data are available on the distribution and occurrence of the 8 GRUMO pesticides in Danish groundwater. A smaller database covers other pesticides and metabolites. Together, these data show that pesticides and metabolites occur particularly in the uppermost and youngest aquifers. In these aquifers, pesticides have been found in 34% of analysed screens in an expanded analytical programme in the interval 0-10 metres below ground level. A planned future expansion of the monitoring system and the water companies’ expanded analytical programme (covering more substances) will show the extent to which groundwater near the surface contains pesticides and metabolites. The frequency of finds of metabolites from the now banned substances atrazine and dichlobenil is high both in the groundwater monitoring system and in the water companies’ raw-water monitoring. It should be noted that the water companies’ raw water often comes from wells in areas near towns, so the finds are characterised by non-agricultural use of pesticides.

In the following, the occurrence of pesticides in watercourses is described. The effects of their occurrence are described in section 5.3. These data have also been used in calibrating the models for calculating the consequences of a partial phase-out of pesticides.

Up to and including 1996, 32 different pesticides were detected in Danish watercourses, corresponding to approx. 30% of the pesticides tested for (see table 4.9). Several different pesticides can be found simultaneously in the individual watercourses. Pesticides are found in watercourses in all seasons, but most frequently in the spraying season and during increased run-off after rain. The frequency of pesticide finds and the concentration of pesticides are generally highest in agricultural catchment areas with clayey soil. Some of the pesticides are found in only a few per cent of the samples, while others are found in up to 64% of them.

The detection frequency is highest in the spraying periods and in connection with precipitation with increased run-off in watercourses. The highest value in the analyses up to and including 1996 was 10 microgramme per litre (bentazon). 12 of the pesticides were found in concentrations above 1 microgramme per litre and 31 substances were found in concentrations above 0.1 microgramme per litre. Glyphosate was tested for in 6 samples and found in all of them in concentrations from 0.02 to 0.21 microgramme per litre (Funen County 1997, Kronvang 1998; Spliid, Mogensen 1995).

Analyses of surface water in Funen County 1997

Since 1996 there have as yet been only a few reports from counties and municipalities. Funen County Council has published the results of analyses of 6 watercourses, 4 springs and 11 drains in agricultural areas (Funen County 1999). Samples were tested for 94 active ingredients and 5 metabolites. Of these, 33 different substances were found in concentrations up to 10 microgrammes per litre (the herbicide bentazon). Of these, 26 are authorised for use today. It should be noted that one of the substances found, metazachlor has never been authorised in Denmark. Most of the substances were found in concentrations above 0.1 microgramme per litre. Up to 18 different pesticides were found in individual watercourses within the same day. The substances found most frequently in 33 samples from watercourses in concentration of or over 0.1 microgramme per litre were the metabolite AMPA from glyphosate (79%), the metabolite BAM from the banned herbicide dichlobenil (48%), isoproturon (36%), glyphosate (31%), the banned herbicide hexazinone (30%), and diuron (24%). The substances and concentrations found largely correspond to earlier finds.

The study indicates that many of the pesticides occurring in the watercourses are transported to them via drains, particularly in the spraying season. It has also been established that there is some transport of pesticides via urban wastewater. Drift does not seem to be of any great significance to the occurrence of pesticides in the watercourses tested. On the basis of observations over a 10-year period, Funen County Council estimates that more than 200 km of watercourse, corresponding to around 20% of the watercourses tested, have been exposed to acute damage, since large numbers of crustaceans and aquatic insects have been killed.

Analyses of water from a watercourse in Kolding Municipality in 1998

In May-June 1998, Kolding Municipal Council tested 14 water samples from Dalby Møllebæk (a mill brook), situated in an agricultural catchment area consisting mainly of moraine clay (Kolding Municipality 1998). The analysis covered 33 pesticides or metabolites, of which 21 were found in concentrations of up to 11 microgramme per litre (simazine). Of these 21 pesticides, 14 are authorised for use today, while 5 are subject to restrictions. The 3 most frequently found substances were BAM, isoproturon and simazine, all of which were found in all the analysed water samples. Also found in at least half of the samples were the substances MCPA, mechlorprop, atrazine, bentazon, desisopropylatrazine, ethofumesate, metamitron, terbuthylazine, bromoxynil, propiconazole, dichlobenil, dichlorprop and ioxynil. The highest concentrations were found in connection with precipitation events. The frequent occurrence of atrazine (79% of the samples) and atrazine’s metabolites indicates that the occurrence came from use of the substance before it was banned in 1994 and thus from the groundwater.

The figures from Funen County for 1997 and those from Kolding Municipality for 1998 will be included in the overall survey for the years in question when the finds are reported. They therefore do not appear in the tables in this chapter as the data supplied are incomplete seen from a countrywide perspective.

Table 4.9
Occurrence of pesticides in Danish watercourses with indication of number of positive finds in relation to number of samples. H = herbicide, F = fungicide, I = insecticide. The samples were collected in the period 1989-1996 (Mogensen 1998).

Substances in italics are banned.
^a = permitted with restricted use/or authorised on special conditions
m g/L = microgramme per litre

Table 4.10 Look here!
Occurrence of pesticides in watercourses, with a breakdown in types of catchment area. The table shows the number of finds out of the total number of samples and the concentration range in microgramme per litre. The first study was carried out by DMU in 1989-1991. In 1994-1996 a major study was carried out in Funen County. In 1996, the table includes data from several counties (Mogensen 1998).

It will be seen from table 4.10 that the frequency of pesticide finds and the concentration of the pesticides are generally highest in agricultural catchment areas with clayey soils. That is presumably due to the fact that farmers on rich soil often farm more intensively and thus spray more often and to the fact that pesticides are quickly transported to watercourses via drains through cracks in the clayey soil. The fact that pesticides have been detected in springs indicates specifically that pesticides can be led to watercourses via groundwater. Fewer different pesticides have been found in forest watercourses, but the herbicides that are used in forests frequently occur in the watercourses. That applies particularly to hexazinone. Many different pesticides have been detected in urban watercourses, which cover both "urban watercourses" and "mixed catchment areas". Agricultural pesticides therefore also occur. Dichlobenil, which is used as a universal herbicide and particularly its metabolite, BAM (2,6-dichlorobenzamide), occur very frequently, both in urban areas, agricultural areas and forest watercourses, see table 4.11. Concerning BAM, readers are referred to section 4.1, where a similar frequency in groundwater is described.

In most of the watercourses, many different pesticides are found at the same time, which is of significance in the assessment of the effects on flora and fauna in the aquatic environment, see section 5.3.

Table 4.11 Look here!
Occurrence of metabolites (of a dichlobenil, b glyphosate, c atrazine) of pesticides in watercourses, with a breakdown by type of catchment area. The table shows the number of finds out of the total number of samples and the concentration range in microgramme per litre. Dichlobenil and atrazine may no longer be used in Denmark. The first studies were carried out by DMU in 1989-1991. In 1994-1996, a major study was carried out in Funen County. In 1996, the table includes data from several counties (Mogensen 1998).

4.2.1 Conclusions

32 different pesticides have been detected in Danish watercourses, corresponding to about 30% of the substances tested for, together with 4 metabolites. The frequency of finds is greatest in the spraying periods and in connection with precipitation events with increased run-off. The highest value of 10 microgrammes per litre was found for bentazon. 12 of the pesticides were found in concentrations above 1 microgramme per litre and 31 substances were found in concentrations above 0.1 microgramme per litre. Glyphosate was tested for in 6 samples and found in all of them in concentrations from 0.02 to 0.21 microgramme per litre (Funen County 1997, Kronvang 1998; Spliid, Mogensen 1995). The pesticides occur in all types of watercourse, but mostly in clayey agricultural catchment areas. Pesticides can be transported to lakes and watercourses with surface run-off. That applies both to water-soluble substances and to substances that bind to particles of soil and that are only transported with heavy run-off, which causes soil erosion. Therefore, both substances that are banned because of their mobility in soil (e.g. atrazine, dichlorprop and hexazinone) and substances that, as far as is known today, are only transported to the groundwater with leaching rainwater under extreme conditions (e.g. glyphosate, esfenvalerate and pirimicarb) have been detected. Also detected is DNOC, which can come from transboundary atmospheric transport and be synthesised from car emissions through atmospheric, chemical processes (see section 4.5).

The studies of watercourses are representative of Denmark

The first study was carried out in the period 1989-91, and the others in the period 1994-97. There are big differences in the sampling intensity and strategy in the various analyses. In some cases, sampling was done on predesignated dates; in some, it was concentrated in the spraying season, and in some, sampling extended over the whole year. Some sampling was done in connection with increased water flow in the watercourse. The analyses did not include samples taken because of suspected pollution from, for example, point sources. The watercourses analysed cover a number of counties and different types of catchment area and types of soil, and must generally be regarded as representative of Denmark. There are not as yet any long time series of measurements of pesticides in watercourses, but monitoring programmes in progress in the counties and countrywide studies will in time result in the necessary database for determining the development.

Pesticides enter lakes, coastal waters and ponds via watercourses, surface run-off from adjacent land, groundwater and atmospheric deposition, including spray drift. The effects of pesticides in stagnant water are described in section 5.3. There are as yet no systematic data on the occurrence of pesticides in Danish lakes and coastal waters. Data from these compartments are therefore not included in the report. In connection with Aquatic Environment Plan II, a measuring programme is being initiated, but results for lakes cannot be expected in the coming year.

The studies from stagnant water cover results from two projects. One includes samples from four ponds in the Køge district (Spliid, Mogensen 1995), while the other includes a number of ponds in South Funen and Avernakø (Briggs in press). In both districts, the ponds are in fields with clayey soil. Many field ponds have neither inflow nor outflow. It must therefore be expected that pesticides remain in them longer than in watercourses and that the risk to the aquatic organisms in them is greater. The results of these studies are given in table 4.12.

The concentration interval in the table indicates the lowest and highest concentration in the samples in which the pesticide in question was detected. Column 3 shows the number of samples analysed for the pesticide in question and the number of samples in which the substance was detected. All the substances for which the samples were analysed were detected, with the exception of bromoxynil and simazine. The highest concentration for an individual substance was 11 microgrammes per litre.

Table 4.12
Pesticides detected in ponds in Danish farmland in studies carried out by DMU in 1989-1991 and by Amphiconsult in 1994-1995 (Mogensen, Spliid 1997)

m g/L = microgramme per litre

4.3.1 Conclusions

There have been only a few analyses of stagnant water, and then only of ponds. Mechlorprop has been detected in concentrations up to 11 microgrammes per litre. In the case of 5 substances, the concentrations exceeded 1.0 microgramme per litre and in the case of 13 of the 15 substances tested for, the concentrations exceeded 0.1 microgramme per litre.

The ponds tested include several in the Køge region and on the island of Avernakø. The analyses were carried out at the beginning of the 1990s. Most of the sampling took place in the spraying season. The analyses of ponds were less extensive than those in watercourses and cannot be regarded as representative of Denmark as a whole.

Soil water can be defined as water that is present in the uppermost part of the soil stratum. Soil water can be collected by installing ceramic or teflon suction cups in the soil to suck water out of the soil matrix. Drain water is water that runs freely through the soil to drainage pipes, which are typically placed at a depth of about 1 metre. Pesticides in drain water can thus, if mobile pesticides are used in periods with a lot of precipitation, have a retention time in the soil of just a few days or weeks. For pesticides that are bound in the soil or that are used in periods without downward movement of water in the soil, the retention time can be months or years.

Soil water samples containing pesticides generally indicate use of the pesticides in question on the surface of the ground directly over the sampling site or transport with rainwater. Pesticides spilled on the surface of the soil or pesticide waste buried near the sampling site can be detected in soil water or drain water, but if several suction cups are established under the same field, a point source occurrence near one of them can be detected.

Drain water samples represent water from the entire catchment area that is drained by the drainage system in question. The catchment area can include many fields with different uses, backfilled marlpits, buried waste, sites for filling and washing of spraying equipment, greenhouses, etc. To be able to interpret the result of a drain water sample it is important to know which catchment area the sample in question represents. By placing vertical drainpipes with screens at a depth of 1 metre one can extract drain water samples that represent the local area around the screen when the soil is saturated.

We have analyses of pesticides in soil and drain water from 10 different localities in Denmark. The samples analysed include samples from Højvands Rende and Bolbro Bæk, which are part of LOOP in Aquatic Environment Plan I. A number of samples have also been taken in special studies and research projects in the last 10 years. However, there have been no systematic analyses of pesticides in soil water and drain water of the kind carried out in groundwater.

The results of all analyses of pesticides and their metabolites in soil and drain water are shown in tables 4.13 and 4.14. Table 4.13 shows the occurrences in fields with sandy soil, and table 4.14 the occurrences in fields with clayey soil. These analyses do not include analyses of drain water from filling and washing sites, greenhouses, and other point sources, where high concentrations of pesticides have been found in special analyses.

The analyses cover 27 pesticides and metabolites. The number of analyses carried out is shown in the tables. The results are given as detections of less than 0.1 microgramme per litre, detections greater than 0.1 microgramme per litre and the highest concentration detected. It will be seen directly from a comparison of tables 4.13 and 4.14 that both the concentration levels and the frequency of detections are highest in localities with clayey soil. The substances detected in the highest concentrations are atrazine (7.8 microgrammes per litre), hexazinone (4.3 microgrammes per litre), dichlorprop (1.4 microgramme per litre) and 2,4-D (1.2 microgramme per litre). Concentrations greater than 0.1 microgramme per litre have also been found for the pesticides isoproturon, bentazon, MCPA, and mechlorprop, and for the degradation products desisopropylatrazine, 2,4-dichlorophenol and ETU.

Some of these detections can be correlated with pesticides used on experimental areas or stated to have been used on the land in question. For example, relatively high concentrations of both atrazine and hexazinone (herbicides) were detected under a Christmas tree plantation, where the substances had previously been used (Felding 1992). In some cases, low concentrations of pesticides have been detected in places where these substances have not been used for years. This could indicate that the retention time for the pesticides in question in the soil has been long in these cases. In a special analysis of the chlorophenoxy acids MCPA, dichlorprop, 2,4-D and mechlorprop, all the substances were detected after 3 years in soil water at a depth of one metre in concentrations that were in several cases greater than 0.1 microgramme per litre (Felding 1993). The detections of ETU are from a trial area treated many times with dithiocarbamate fungicides, which break down into ETU (Spliid 1998a).

As a test system between full-scale field analyses and simple laboratory analyses, lysimeter analyses with undisturbed columns of soil can be used to predict whether a pesticide or its degradation products can leach from a column of soil. Such lysimeter analyses now form part of the documentation material used as the basis for approving new pesticides. A lysimeter can consist of a block of soil sampled at the site in a steel frame without being disturbed. The lysimeter is moved to the test locality, where crops are grown in it and it is treated with the pesticide in the ordinary way. The naturally or artificially supplied water passing the block of soil is collected. Using a radioactively labelled pesticide, one can determine whether there is any breakthrough of radioactivity and thus of pesticide or degradation products. Where possible, the substances passing the soil column are identified by means of chromatographic methods and by comparison with reference substances. A lysimeter analysis is typically carried out several years after the pesticide has been applied. It can be used to determine the risk of a pesticide leaching in different types of soil, with different cultivation conditions and in different precipitation situations, with a surface area from 0.25 to 1 m².

To determine the mobility in columns of soil in the deeper soil layers, columns with a surface area of, for example, 0.25 m² can be sampled in a steel cylinder and taken to the laboratory, where the temperature and the groundwater’s movements can be simulated and controlled.

Table 4.13
Pesticides and metabolites detected in fields with sandy soil (Spliid 1998a)

n.d. = not detected
^M = metabolites
Italics = substances authorised in 1998, possibly with dispensation
m g/L = microgramme per litre

Table 4.14
Pesticides and metabolites detected in fields with clayey soil (Spliid 1998a)

n.d. = not detected
^M = metabolites
Italics = substances authorised in 1998, possibly with dispensation
m g/L = microgramme per litre

4.4.1 Conclusions

In a number of analyses, pesticides have been detected in soil water and drain water from fields that have been treated with pesticides. In the analyses described, there is information concerning the use of pesticides on the fields in several projects, while for a few localities, the figures cover the catchment area in general, and it is not possible to demonstrate a relationship between dosage and detections for the individual field. With the available data it is not deemed warranted to draw conclusions concerning relationships between dosage and concentration of pesticides in the water samples analysed.

As found by Felding (1998a), there have been relatively few studies in Denmark. In a cooperative project under the auspices of the Nordic Council of Ministers, rainwater samples were collected from 2 localities in Denmark – Ulfborg Plantation 10 km from the west coast of Jutland, and Gadevang near Gribskov, a forest in North Zealand – from 1992 to 1994. The analyses covered 10 pesticides: propiconazole, prochloraz, lambda-cyhalothrin, cypermethrin, esfenvalerate, deltamethrin, atrazine, mechlorprop, dichlorprop and MCPA. Only the phenoxy acids were detected, and the highest concentrations found were just under 0.4 microgramme per litre (Kirknel et al. 1997). The concentrations found were very small in relation to the dosage in field spraying. The effects of pesticides in rainwater are assessed in section 5.2.

In a current project, samples of precipitation are being collected from three Zealand localities: Gadevang, Gisselfeld and Lorup. The samples are analysed for the phenoxy acids: MCPA, mechlorprop and dichlorprop, and for the herbicide isoproturon. The highest concentrations found so far are just over 0.6 microgramme per litre for the phenoxy acids and just under 0.4 microgramme per litre for isoproturon. In most cases, the time when the herbicides were detected in the precipitation coincided with the time of use (Felding 1998a).

In autumn 1997, 13 mixed samples extracted from the above-mentioned 3 localities from September 1996 to November 1997 were analysed for 44 pesticide chemicals. Table 4.15 shows the content of the 13 samples. In all, 8 pesticide chemicals were detected: isoproturon, metamitron, DNOC, mechlorprop, methabenzthiazuron, 2-hydroxyterbuthylazine, terbuthylazine and 2,4-D. DNOC was found to be present during the entire period in a relatively high concentration range from 0.3 to 4.5 microgramme per litre. This substance has not been used in Denmark for the last 10 years. The substance can also be formed by atmospheric, chemical reactions, see section 4.5. There may thus be transboundary air pollution. DNOC has also been detected in groundwater near the surface and in watercourses (Spliid et al. 1996), see sections 4.1.1 and 4.1.2.

In 1990-1991, DMU measured the content of a -HCH and g -HCH (lindane) in rainwater in 2 localities in Denmark: Husby and Ulfborg in West Jutland. In 1992, analyses were carried out of samples from 3 localities: Ulborg, Bagenkorp and Anholt. The maximum concentration was found to be 0.1 microgramme per litre. The analysis indicated that the occurrence of lindane came from use in countries south and west of Denmark (Cleemann et al. 1995).

Table 4.15 Look here!
Pesticides and degradation products in rainwater samples collected in the period September 1996 to November 1997. In all, 13 samples were collected over a period of 2-3 weeks in Lorup, Gisselfeld and Gadevang on Zealand. DMU analysed the samples for 36 active ingredients and 8 metabolites. The concentrations are given in microgramme per litre. It will be seen that for isoproturon and DNOC, the concentration exceeds the drinking water limit of 0.1 microgramme per litre (marked in bold) (Felding 1998a).

Table 4.16 Look here!
Pesticides and degradation products in rainwater samples from Zealand. The table shows the figures from table 4.15 for the deposition of pesticides, calculated as g per hectare (Felding 1998a). The analysis covers 36 active ingredients and 8 metabolites.

In table 4.16, the measured concentration of pesticide chemicals from table 4.15 is converted from microgramme per litre to gramme per ha. In this analysis, which does not cover all volatile pesticides, DNOC (4,6-dinitro-2-methylphenol) constitutes just under 90% of the total quantity of pesticide chemicals in the atmospheric deposition with rainwater. The distribution of phenols and nitrophenols in clouds and the occurrence and formation of phenols in the atmosphere have been described by Lüttke and Levsen (1997) and Lüttke et al. (1997). It is concluded in these studies that DNOC occurs mainly in gaseous form rather than in liquid form in the atmosphere. Dinitrophenols are mainly formed through reactions in the atmosphere, unlike mononitrophenols, which are primarily formed in connection with exhaust fumes from cars.

Pesticide chemicals have been detected in precipitation collected in Denmark. The study was limited and covered only a few substances and not those with the highest potential for volatilisation. In most cases, there was a relationship between the spraying season and the time of detection. However, pesticides that are no longer used in Denmark were also detected. These substances were presumably transported here over long distances or originated from other sources than treatment with pesticides in agriculture. The prohibited substances found in the precipitation included DNOC, which was found in rainwater throughout the year and in by far the highest concentrations among the substances analysed for. It is therefore highly likely that its presence is due primarily to the formation of nitrophenols in the atmosphere. The international literature describes the formation of nitrophenols in the atmosphere as a reaction between benzene, toluene and NO_x. The exhaust fumes from cars contain mononitrophenols and other nitrophenols, which contribute to the atmosphere’s content of these substances. However, the use of DNOC as a pesticide outside Denmark may also contribute to the content of nitrophenol in the atmosphere.

Pesticides can be transported with water running on the surface of the ground. Surface run-off from sloping fields can carry both water-soluble substances and substances that are adsorbed to particles of soil out into watercourses and lakes. Particle-borne transport occurs during heavy run-off of rainwater or melt water and soil erosion.

In a project in the Danish Environmental Protection Agency’s Pesticide Research Programme (Felding et al. 1997), the surface run-off of two relatively water-soluble herbicides (mechlorprop and dichlorprop) and a sparingly water-soluble insecticide (alfacypermethrin) was studied during two growth seasons (1992-1993). The field had an average gradient of 12% and was used for winter wheat in both years.

Table 4.17 shows the quantities of pesticide transported with surface run-off in the years mentioned. For both years, the quantities, stated in parts per thousand of the sprayed pesticide, were 0.08 for mechlorprop, 0.002 for dichlorprop and 0.001 for alphacypermethrin.

Mechlorprop and dichlorprop were detected particularly in the aqueous phase of the surface run-off, while alphacypermethrin was only detected in samples of water containing particles of soil, which accords with the fact that this substance adsorbs strongly to soil particles. It will be seen that the biggest run-off was of substances applied in the autumn.

Table 4.17 Look here!
Spraying of pesticides on the trial field and pesticides detected in surface run-off (Felding et al. 1997). The substances in question were authorised at the time of spraying.

Events with run-off occur only, and momentarily, when the precipitation within 24 hours exceeds 10 mm (DMU 1995); Groenendijk et al. 1994; Liess et al. 1999; Møhlenberg, Gustavson 1999). On average, this occurs three times a year in Denmark (Funen County, South Jutland County). During precipitation events of more than 10 mm, the surface run-off of pesticides make up 0.2 % of the pesticide pool from the nearest 2 ha in the field, which, according to Groenendijk et al 1994, is a good estimate in countryside with slight gradients. Swedish studies from 1990-1996 (Kreuger 1998; Kreuger, Tornqvist 1998) estimate correspondingly that 0.1-0.3% of the pesticide spread on fields in the catchment area are lost to the aquatic environment. Recent German studies carried out near Braunschweig, which has the same type of soil and field topography as Eastern Denmark, has shown that the run-off of fenvalerate, among other pesticides, occurred in pulses and was clearly associated with precipitation events of more than 10 mm per day (Liess et al. 1999). The total transport to a watercourse was calculated at between 0.012% and 0.068% of the total amount applied to a 9 ha field. Converting to a run-off area of 2 ha and taking account of the degradation as used in the model in section 10.3.3, that corresponds to a loss of 0.05-0.03% with surface run-off (Møhlenberg, Gustavson 1999). Møhlenberg, Gustavson use a loss of 0.2% from 2 ha as a conservative estimate in the model calculations in section 10.3.3.

The main measure used to try to prevent pollution of watercourses and lakes is the establishment of buffer zones along watercourses and lakes within which cultivation is not allowed. The buffer zone acts as a filter and reduces the amount of surface run-off.

4.6.2 Conclusions

"Spray drift" should be understood to mean the amount of pesticide that is not deposited on the field being sprayed. Pesticides can be transported via the atmosphere to land outside the sprayed area in the following two ways:

During field spraying, the spray liquid is relatively finely atomised in order to achieve a uniform distribution on the area and good deposition of the spray liquid on the crop and pests. During the droplets’ transport to the crop and its pests, loss can occur through the droplets being carried out of the sprayed area in particulate form as drift or in the form of vapour. The loss can occur both on the droplet’s way to the target and from the target of the spraying for a period after spraying. Drift, together with volatilisation of pesticides, explains why pesticides are detected in precipitation, surface water and on unsprayed areas. Drift is the reason why, with very low vapour pressure, pesticides can be measured in the atmosphere and in precipitation, because these substances can be transported as small particles that form from drops of spray fluid that are borne and dried in the air.

Droplet size, volatilisation and relative humidity

Volatilisation from such small droplets simply reduces the size of the droplets still further and produces very small droplets that are transported with wind over long distances. Droplets with a diameter of less than 50 micrometres have a critical size because they remain suspended in the air for a relatively long time. If the relative humidity is low, large droplets can also become smaller, whereby the risk of drift increases considerably. It has thus been calculated that the proportion of air-borne droplets at a distance of 500 metres from the sprayed areas increases more than tenfold if the relative humidity falls from 100% to 50% at 20^oC (Thomson, Ley 1982).

Height of spraying boom above the ground

The height of the spraying boom has a considerable effect on drift. It is primarily the small droplets – 100 micrometres or less – that are affected by changes in the height of the boom. Drift thus doubles when the boom height is increased from 50 cm to 70 cm with a traditional flat-spray nozzle (Miller 1988).

As mentioned in section 9.6, spraying equipment is undergoing further development that could help to reduce drift.

4.6.4 Conclusions

Drift to the surrounding areas implies a risk of exposure of hedgerows, dykes, dry stone walls and small biotopes in farmland and of natural terrestrial and aquatic areas. Drift, together with volatilisation of pesticides, explains why pesticides are detected in precipitation, surface water and on unsprayed areas. Drift depends particularly on the droplet size and wind velocity. The droplet size depends on the spraying equipment and spraying technique used.

Pesticides evaporate both during and, especially, after spraying. Together with drift, see section 4.6.2, volatilisation conveys relatively large quantities of pesticides to the atmosphere, which is thus, quantitatively, the principal transport path for pesticides away from the sprayed area. Lastly, it should be noted that considerable quantities of pesticides can be transported by wind erosion – for example, if there is a storm after spraying of crops in the spring, when the plant density is low. The volatilisation depends on the properties of the substance and, particularly, on its vapour pressure. The temperature, water solubility and adsorption to the soil and plant surfaces, the humidity of the soil, atmospheric turbulence, and the concentration of the pesticide, including its degradation, also play an important part. It is difficult to measure the volatilisation because of its great natural variability (Løkke 1998).

Models for volatilisation

Volatilisation of pesticides depends on the aforementioned factors and can be calculated by means of mathematical models that include the main factors governing the volatilisation. Jansma and Linders (1995) calculated the volatilisation from the surface of the earth by means of the so-called "Dow method", which takes account of the pesticides’ vapour pressure, water solubility and adsorption to soil as the factors governing the volatilisation. The method was validated by comparison with measured values. This method can thus be used for crops grown in rows, where the plant cover is small and most of the pesticides hit the surface of the earth. The calculated values generally lay within a factor of 7 from the measured values. However, this model has a tendency to overestimate the volatilisation. There are as yet no suitable models for estimating the volatilisation of pesticides that are mixed with the soil during harrowing and ploughing. It is also difficult to make a general model for the volatilisation of pesticides from the surface of the plants.

Volatilisation from soil

In the case of many pesticides, most of the pesticide that lands on the ground can in theory evaporate within a few days, depending on the climatic conditions. Volatilisation increases with rising temperature and wind velocity. If the relative humidity is low, the ground surface dries out, which reduces volatilisation.

Volatilisation from plants

Little is known about volatilisation of pesticides from plants. High relative humidity increases the volatilisation but can at the same time increase the absorption of a pesticide by the plant. Once the pesticide has penetrated the plant, 95% of it remains there. The volatilisation depends on the species of plant, the amount of foliage per unit of area and the nature of the surfaces of the plant. The spraying technique and the coformulants in the pesticide formulation also play a part. These substances are intended to ensure that pesticides are spread and attach to and penetrate the plant. The ancillary substances can thus help reduce volatilisation and thus increase the substances action time in and on the plant. Volatilisation is greatest from small droplets, which have the relatively largest surface, and least from large droplets, partly because absorption by the plant is greatest from large droplets. The Dutch authorities have estimated that about 20% of a sprayed pesticide evaporates from plant surfaces and is transported to the atmosphere (Ministerie van L.N.V. 1991).

Measurements of volatilisation in the laboratory and the field

There have been many foreign studies of the volatilisation of pesticides. Most of them were carried out more than 10 years ago with pesticides that are no longer used in Denmark. Jansma and Linders (1995) have reviewed data from the literature. It appears from their work that both the herbicide DNOC and the chlorinated insecticide lindane, which have been detected in rainwater in Denmark (see section 4.1.5), evaporate very quickly from plant surfaces. In German field tests with different crops, it was found, for example, that 45% of lindane evaporated from green beans within one hour of spraying, 50% from lettuce, 30% from kohlrabi and 25% from spring wheat (Boehncke et al.1990). After 3.1 days, 88% had evaporated from the spring wheat and more than 90% from the other crops. In several studies, measurements of the volatilisation of lindane from the ground showed a lower rate of volatilisation but still so much that most of the sprayed pesticide was transported to the atmosphere.

Measurements exist of the volatilisation of individual substances included in measurements of the content of pesticides in Danish rainwater. The substances include atrazine and simazine.

For atrazine, measurements in field studies in the USA have shown that up to 9% can evaporate from the ground surface in 35 days. In another study, 1.3% of simazine evaporated from the ground within 21 days. Measurements of volatilisation from the ground of substances now in use in Northern Europe, showed up to 49% of chlorpyrifos after 26 days, up to 52% of deltamethrin after 3.1 days and 90% of trifluralin after 2.5-7 days. Field tests in Germany of volatilisation of the insecticide deltamethrin after 3.1 days showed 72% from green beans, 34% from kohlrabi, 70% from lettuce and 24% from spring wheat (Boehncke et al. 1990). The same authors measured up to 100% volatilisation of mevinphos from plants after 3.1 days.

Calculation of the volatilisation from the surface of the ground

The simplest model for calculating volatilisation is a so-called 1st order model. In the "Dow model" it is assumed that the so-called rate constant for volatilisation is directly proportional to the vapour pressure and inversely proportional to the water solubility and the soil-adsorption constant (Jansma, Linders 1995). This model is used in the EU´s PC-based expert system for assessment of chemical substances, EUSES (EC 1996). The volatilisation according to this model is shown in table 4.18 for substances with high volatilisation that are also widely used. However, the model overestimates the volatilisation. Another model, developed recently in the Netherlands, includes the substances’ fate in the atmosphere, deposition and effects on plants. The results indicate that, with the pattern of use in the Netherlands in the period 1985-1995, in all 5.5% of the amount of herbicides applied evaporated. The atmospheric deposition corresponded to an average treatment frequency index on nature areas of 0.02 per year (Klepper et al. 1998).

Table 4.18
Calculated volatilisation of pesticides with potential volatilisation from ground surfaces (Jansma, Linders 1995; Løkke 1998).

However, these calculations also showed that there is very little volatilisation from ground surfaces of most of the pesticides used in Denmark today. There is apparently considerable volatilisation of fenpropimorph, pendimethalin, permethrin, propyzamide and prosulfocarb. There is thus a risk of these substances being spread in the atmosphere and detected in rainwater and surface water. The fate and occurrence of these substances in the atmosphere over Denmark have not been investigated and are not covered by DMU’s analytical programme for rainwater (Løkke 1998).

4.6.6 Conclusions

Pesticides evaporate both during and, especially, after spraying. Together with drift, see section 4.6.2, and occasionally wind erosion of soil treated with pesticides, volatilisation transports relatively large quantities of pesticides to the atmosphere and is thus, quantitatively, the principal transport path for pesticides away from the sprayed area. The volatilisation depends on the properties of the substance and, particularly, on its vapour pressure. The temperature, water solubility and adsorption to the soil and plant surfaces, the humidity of the soil, atmospheric turbulence, and the concentration of the pesticide, including its degradation, also play an important part. It is possible to calculate the volatilisation from ground surfaces by means of simple mathematical models, whereas there are as yet no similar models for calculating the volatilisation from plant surfaces. According to model calculations, there can be considerable volatilisation from ground surfaces – theoretically right up to 100% within a few days in the case of some substances. In the Netherlands it is estimated that about 20% of the total amount placed of all types of pesticides evaporates.

Model calculations show that very little of most of the pesticides used in Denmark is transported from ground surfaces to the atmosphere through volatilisation, although some substances show a big potential for volatilisation.

Pesticides degrade in a stepwise process, often via the formation of metabolites, the chemical structure of which may be similar to the original pesticide. A pesticide’s metabolites can be more toxic and leach more easily than the original pesticides. This applies, for example, to the atrazine metabolites, deethylatrazine and desisopropylatrazine. It is therefore of great importance that studies of the persistence of pesticides in the soil and leaching from the soil include corresponding studies of metabolites (Fomsgaard 1998).

Degradation and sorption are key factors in the assessment of pesticides’ persistence in the soil. The time of application and the locality also affect the persistence of pesticides. The amount of organic matter in the soil can lead to greater sorption and thus slower degradation and slower leaching. Conversely, the organic matter can also increase the presence of microorganisms, which can in turn increase the rate of degradation. Extreme soil environments, such as the surface on railway lines, paths and car parks, are such poor environments for biological activity that herbicides used on these areas are degraded very slowly. Furthermore since these types of soil bind pesticides poorly, they have a high potential for pollution of groundwater.

Processes that remove pesticides from the soil

A wide range of processes – surface run-off, volatilisation, uptake in plants, leaching and degradation – help to remove pesticide residues from the soil after spraying. Degradation and metabolism in plants differ from the other processes by transforming the substance or removing it permanently from the environment. Surface run-off and leaching can carry the substance into surface water or groundwater; volatilisation can lead to the occurrence of pesticide residues in rainwater; and uptake by plants can lead to contaminated food if the pesticide is not broken down after uptake. Pesticide residues can also be bound in the soil without being degraded.

Persistent pesticides

Pesticides are generally either degraded abiotically, e.g. by photodegradation, or biologically, primarily through biological conversion by microorganisms. A persistent pesticide is characterised by remaining intact in its active form if it is not sensitive to physical and chemical factors or if the microorganisms cannot degrade it. In addition, metabolites, degradation products or reaction products that can be similarly persistent can be formed. The pesticide and/or these metabolites can be inaccessible to degradation if they are adsorbed on humus or clay minerals. The distribution of pesticides between the soil’s pore water and particles of soil is an equilibrium between free molecules and molecules bound to either dissolved organic matter or to attached soil particles. There is therefore a possibility that pesticides bound in the soil can later be released as a consequence of a change in the physical-chemical environment, e.g. acidification of the soil or degradation of humus or clay minerals.

Persistence analyses in the authorisation of pesticides

In connection with the authorisation of pesticides, laboratory tests are carried out as a basis for assessing the persistence. The active ingredient or its metabolites, degradation products or reaction products that are of environmental importance must not have a half-life of more than 3 months and not more than 10% of the amount applied must remain after one year. If only a small proportion of the active ingredient (i.e. less than 5% in 100 days) is completely broken down into CO₂, water and salts, there is a risk of it being bound unchanged in the soil. It is therefore a requirement that not more than 50% of the active ingredient may be bound in the soil after 30 days or max. 70% after 100 days. If these values are exceeded, an evaluation of the conditions for use is carried out and relevant field trials may also be included in the assessment. Active ingredients that have half-lives of more than 6 months and that imply a risk of exposure of the external environment are regarded as possessing unacceptable persistence. This means that the pesticide in question cannot be authorised.

In cases in which relatively large amounts of a persistent pesticide are bound in the soil after repeated application without being directly accessible, it is known that small amounts of the pesticide can be absorbed by plants, earthworms and microorganisms. With continuous accumulation of persistent pesticides in the soil, this absorption must be expected to increase. It is not known how such absorption will affect organisms in the soil in the longer term (the so-called chronic effects). In cases in which a pesticide is bound very strongly to the soil without breaking down, it will not be possible within the foreseeable future to carry out long-term trials to see whether the ingredient and its relevant metabolites, degradation products and reaction products are released and cause damage to the soil’s ecosystem. In such cases, the authorities may decide that an active ingredient must not be authorised due to unacceptable persistence. This authorisation practice is used in Denmark and the Netherlands, and Denmark is seeking to get it implemented in the EU.

Soil treatment

Soil treatment affects the chemical, physical and biological factors in the soil and thus has a major indirect effect on the persistence and leaching of pesticides. In the international literature, the relationship between leaching and soil treatment has been studied by Hatfield (1996), Locke and Harper (1991) and Gaston et al. (1996). If soil treatment is increased, some of the macropores would be destroyed. That means that the pesticides’ retention time would probably increase in the ploughing layer, where the potential for degradation is greatest, and leaching would be reduced. On the other hand, surface run-off might increase. If soil treatment were reduced or eliminated, transport in macropores would increase and thus also leaching of pesticides. Reduced soil treatment can also have the opposite effect, since volatilisation increases when the soil is not treated.

Metabolites

The degradation of a pesticide is a phased process that often proceeds via the formation of metabolites, the chemical structure of which is often similar to that of the original pesticide. Complete mineralisation leads to the formation of CO₂, salts and H₂O. Only some of the amount of pesticides that may end up in the soil after spraying will be fully mineralised (i.e. converted into water, CO₂ and salts). The rest of it will be leached or strongly bound or chemically incorporated in the soil’s humus and/or biomass. Both mineralisation and strong adsorption (non-desorbable) or chemical incorporation in, for example, humus remove the environmentally harmful effects of the pesticide/metabolites. The metabolites formed by degradation of pesticides can be not only more toxic to the environment but also more water-soluble and often leach more easily than the original pesticide. If the metabolite is more toxic or leaches more easily, it can present a bigger problem than the original pesticide. The detection of metabolites of dichlobenil and atrazine in groundwater illustrates this.

In the national groundwater monitoring programme (GEUS 1997), analyses were carried out in many Danish counties in 1997 for the metabolite 2,6-dichlorobenzamide (BAM), which is the degradation product of the herbicide dichlobenil, and for atrazine’s three primary degradation products deethylatrazine, desisopropylatrazine and hydroxyatrazine (see section 4.1). The occurrence in groundwater is an important indicator that future studies of the persistence of pesticides in the soil should include the persistence of metabolites with high water solubility and/or low sorption. The sorption can be expressed by the adsorption constant K_d. The value of K_d usually increases with the soil’s content of organic matter. The adsorption constant and solubility of atrazine and the three primary metabolites are shown in table 4.19. Deethylatrazine and desisopropylatrazine are less adsorbable than atrazine.

Table 4.19  Look here!
Water solubility and adsorption constants (K_d) for atrazine and primary atrazine metabolites. Atrazine is banned in Denmark.

DEPA has published a list of pesticides currently used in Denmark and the associated metabolites (DEPA 1997). Selected examples from this list are shown in table 4.20. It is just as important to determine the persistence of metabolites as the persistence of the pesticides themselves because, as mentioned, the metabolites are in some cases more toxic than the pesticides from which they are formed.

Table 4.20
Pesticides with associated metabolites. Selected from DEPA (1997).

Relationship between degradation and adsorption, transport, spraying time and locality

Most pesticides that are strongly bound in soil degrade slowly. This is because the adsorption reduces the contact between the pesticides and the microorganisms carrying out the degradation. The pesticides can be released again (desorbed). The rates of adsorption and desorption play an important role in leaching. The greater the adsorption and the slower the desorption, the lower the probability that the substances will be leached through the soil to the groundwater.

The presence of organic matter, on the other hand, can increase the rate of degradation if the organic matter supplies nutrition and thus growth for the microorganisms that break down the pesticide. This has been demonstrated by Mueller et al. (1992) for fluometuron, by Veeh et al. (1996) for 2,4-D and by Walker et al. (1983) for simazine. The amount of organic matter in the soil at the individual locality thus affects the rate of degradation, although it is not known exactly how.

The temperature can affect the metabolism of many herbicides. For example, on average, the rate of degradation is reduced 2.2 times with a temperature reduction from 20°C to 10°C (Walker et al. 1996). Metabolism of pesticides therefore takes place far more slowly in the wintertime. Application in the autumn with subsequent low temperatures and percolation of precipitation therefore has a much greater potential for pollution of groundwater than application of the same substance in the spring or summer.

4.6.8 Conclusions

The rate of degradation and the adsorption of pesticides are key factors in the assessment of the pesticides’ persistence in the soil. The rate of degradation is affected by a number of environmental factors. The effect of the temperature, the depth of the soil and the water content are of particular interest. Lower temperature and greater depth of soil generally reduce the rate of degradation. A very low water content reduces the possibilities of contact between the pesticide and microorganisms. A very high water content often reduces the rate of degradation because the supply of oxygen can be reduced. In recent studies it has been found that many other factors also influence the degradation of pesticides. For example, the initial concentration of a pesticide greatly affects the rate of degradation. With a high concentration of a pesticide, the degradation process is very slow.

A pesticide’s metabolites can be more toxic and leach more easily than the original pesticides. This applies, for example, to the atrazine metabolites, deethylatrazine and desisopropylatrazine. It is therefore very important for studies of the persistence of pesticides in the soil and leaching from the soil to include corresponding studies of metabolites.

Extreme soil environments, such as the surface on railway lines, paths and car parks, are such poor environments for biological activity that herbicides used on these areas are degraded very slowly. Furthermore since these types of soil bind pesticides poorly, they have a high potential for pollution of groundwater. For many substances, application in the autumn, with reduced temperature and bigger precipitation, can involve a higher risk of groundwater pollution than application in the summer and spring.

No data are available on the extent of washing and filling sites as point sources of pollution of groundwater and surface water. There is thus no documentation showing whether this is a real problem, although there are indications that point sources with pesticides do pollute. For example, pesticides are sometimes detected in groundwater or watercourses in such high concentrations that it is unlikely that the source is application on fields. Point-source pollution can thus explain some local instances found today of limit values having been exceeded in groundwater and wells.

In one of the few studies carried out, Jørgensen and Spliid (1993) examined a washing and filling site in an orchard and found a very large content of phenoxy acids in the analysed soil samples.

Other studies have shown that farmers’ wells and boreholes can be seriously contaminated with pesticides (Anon 1995; Spliid 1998b) caused either by point sources or by use of herbicides on farmyard areas. Areas surfaced with gravel and stone present a special risk of leaching of pesticides compared with ordinary agricultural land. Areas of this type are found in farmyards (Jacobsen et al. 1998).

A German study by Fischer et al. (1996) showed that 98% of the considerable pollution of a watercourse was caused by discharge of isoproturon directly from yard areas via sewerage systems or drains.

Potential sources of pesticides in groundwater

Where pesticides are used and handled outside normal field spraying, there is a risk of pollution of the surrounding environment – of the groundwater, the farmer’s own well or borehole and watercourses – via drainage systems from the farm yard. Table 4.21 shows the potential water pollution from filling and washing sites.

There are no data showing whether washing water or pesticide spills cause serious groundwater pollution, nor is it known whether buried packaging is a serious problem. Before 1980. the authorities recommended burying empty packaging.

In directions from the Pesticide Board of the former Ministry of Agriculture from 1966 it was stated in a section entitled "Disposal of residual stocks of pesticides" that small remnants of pesticides and up to approx. 1 kg should be emptied into an 0.5 metre deep hole and that the packaging should be destroyed. The hole had to be at least 50 m from any well, watercourse, lake or drainage pipes. In the case of larger quantities, it was stated that these should be buried at a tip after permission for this had been obtained from the local health inspector.

Table 4.21
Potential sources of water pollution from filling and washing sites (Jensen et al. 1998).

IPU = Isoproturon.

A number of factors mean that the areas mentioned can be particularly critical. Pesticides from buried packaging and from filling and washing sites can affect the surroundings in a very concentrated form. They can also impede microbial activity, bringing degradation more or less to a standstill. Washing and filling sites are used regularly for many years. That means a big area load, compared with use of the pesticide for field spraying. In addition, washing and filling sites are often unsuitably sited on areas surfaced with gravel and stone but without any organic topsoil. This significantly increases the risk of pollution. When topsoil, with its content of humus, is removed, so are the microorganisms on which biological degradation of spilled pesticides mainly depends. And the big potential for sorption and retention is removed as well. That results in a relatively high rate of transport of water and pesticides. For the same reason, the use of herbicides in farmyard areas and lanes puts well water and groundwater at risk.

Another reason why pollution from filling and washing sites is particularly problematical is that the pesticides are led to the ground with a relatively large quantity of water, which increases the risk of leaching. The yards may also be connected with drains, so the pesticides can be led to watercourses or lakes. In addition, spraying equipment is often washed and filled close to wells, so leaching can occur directly in them unless special measures are taken to prevent it.

Table 4.22
Examples of pesticides in water used to wash spraying equipment. As the examples show, emptying even the diluted washing water out onto the ground can present a risk. The spray-liquid tank should be washed out several times with even small amounts of water. As the examples show, diluting 10 litres twice to 100 litres gives 100 times dilution. If the 10 litres are diluted to 50 litres in four goes, the water consumption will be only 160 litres, but the dilution will now be 625 times greater (Jensen et al. 1998).

IPU = isoproturon.
m g/L = microgramme per litre

A model calculation by GEUS of the potential importance of point sources to groundwater pollution shows that only a minimal proportion of pesticide detections comes from point sources (GEUS 1996). The calculations are based on various assumptions – for example, that the size of the point source is only 0.3 x 0.3 m. Since washing sites and sprayed farmyard areas are usually much larger than that, the leaching could be much greater than calculated.

In an ongoing project in which DIAS, GEUS and DMU are participating, the extent of a pesticide plume from a washing and filling site is being studied. This will provide supplementary knowledge about the importance of point sources.

4.6.10 Conclusions

4.7 The sub-committee’s conclusions and recommendations

1. There are studies of pesticides in the different compartments – groundwater, watercourses, drain water, soil water and rainwater. There are only a few measurements of pesticides in ponds and lakes. Only in the case of groundwater are there time series, but the measuring programmes have not been going on long enough to enable a description of the trends.
2. In the expanded analyses in the groundwater monitoring programme, pesticides or degradation products were detected in 34% of the screens examined in the interval 0 to 10 metres below soil surface. The limit value was exceeded in 23% of the screens. The detection frequency decreases with the depth, which may indicate that remnants of the last 50 years’ increasing use of pesticides are moving towards the deeper aquifers, with increasing future groundwater pollution as a consequence. Another possible explanation might be that degradation is taking place during the onward movement because the concentrations in the deeper soil strata have been exposed to biological and chemical degradation for a longer period of time than concentrations in the uppermost strata. Only when sufficiently long time series from the monitoring programmes are available – which means in 5 to 10 years time – will it be possible to test these hypotheses. The latest research indicates that the degradation of certain pesticides in the aquifers is very slow, while others show degradation.
3. In the expanded analyses in the groundwater monitoring programme, pesticides or degradation products were found in 21% of the screens examined. The limit value was exceeded in 13% of the screens. The degradation product BAM from the now banned herbicide dichlobenil was detected in about 30% of the wells analysed in connection with the water companies’ raw water monitoring system. However, a wide range of substances that are used in agriculture for treating crops are also present in relatively many wells analysed in connection with the water companies’ raw water monitoring system. Larger amounts of pesticides were detected in drain water and soil water than in groundwater, and this may reflect concentrations that may similarly later move towards the groundwater, during which they may undergo degradation and possible formation of metabolites. In both watercourses and ponds, a number of pesticides have been detected in higher concentrations than the concentrations that have effected aquatic organisms in the laboratory.
4. It is a condition of authorisation of pesticides that they degrade in the environment and are metabolised into water, carbon dioxide and salts or into harmless organic substances. However, the sub-committee has noted that, in the case of pesticides detected in the different compartments, it is often only a fraction of the quantities used that can be accounted for, apart from degradation. There is thus a lack of information concerning the total mass flows and the largest flows, including volatilisation and drift, and a lack of actual systematic measurements in the environment of degradation and metabolism as an element of the integrated mass flow analysis. It is thus not possible to make a real and complete description of the fate of the pesticides in relation to the environmental and health loads.
5. The detection of authorised pesticides above the limit value in both groundwater near the surface and in aquifers at greater depths indicates that the present authorisation scheme does not ensure completely against future pollution of the groundwater. As a safety measure, a warning system for pesticides has been established with a view to enabling fast evaluation and, if necessary, removal of authorised pesticides that are a threat to the groundwater.

At present, authorisation of pesticides in Denmark and the EU is based on analyses of research results and assessment of the consequences for health and the environment. This means, in particular, that the fate of the pesticides is not subject to an analysis of the uncertainties and actual variations that form part of an integrated mass flow analysis – in part because such an analysis would require data concerning the actual use, dispersal and degradation under Danish conditions. The sub-committee recommends that an actual mass flow analysis be carried out when pesticides are being reviewed for renewal of their authorisation in pursuance of section 33(4) of the Act on Chemical Substances and Products. This analysis must include both average and "worst-case" scenarios based on measurements taken and experience gained in the period in which the substance in question has been used. If, in this mass flow analysis, knowledge is lacking concerning the individual flows, the sub-committee recommends that the precautionary principle be applied in the assessment of the substances in order to counteract or prevent any consequences of the dispersal of the pesticides and exposure of people and the environment.

The sub-committee recommends that the gross list drawn up with a view to reexamining leaching of pesticides be included in recommendations concerning substitution with less dangerous substances. The sub-committee also recommends that new substances be assessed in relation to the gross list and in relation to non-chemical, alternative methods.

The sub-committee also recommends that volatilisation and chemical reactions of pesticides in the atmosphere be taken into account in connection with authorisation of the pesticides.

5. Environmental impacts

This section describes the known direct and indirect impacts of pesticides on flora and fauna in terrestrial and aquatic ecosystems. The main impacts occur during the application of the pesticides, when organisms are directly hit. However, indirect impacts, which occur as a consequence of the effect on food chains, can also be substantial. Bioaccumulation of pesticides in organisms and possible concentration in food chains are not covered by this report because pesticides with bioaccumulability are not authorised in Denmark.

During spraying, there is spray drift to the surrounding areas. However, hedgerows, dikes, dry stone walls and other small biotopes are of such small width that they should in practice be included in the area that is affected by spray products. Spray drift can affect both terrestrial and aquatic ecosystems. Among the aquatic ecosystems, it is particularly ponds, watercourses and lakes near fields that could potentially be affected. Seen overall, it is not the individual field and its possible loss of wild flora that are the problem but, rather, the countrywide impact on the characteristic farmland flora. Putting it in popular terms, small, uncultivated biotopes like ponds, hedges, dikes and hedgerows can be regarded as small oases in large expanses of monocultures. Large distances between these oases reduce the dispersal and remigration of species and therefore increase the risk of local eradication. It has become more difficult for some bird species to find nesting places in the large, uniform fields. If there is too little food in one type of crop, it is often also too far to fly to the next. The impact of pesticides must thus also be seen in relation to the physical structure of cultivated land and other impact factors, such as fertilisation and crop rotations.

5.1 Impact of pesticides on the fauna in cultivated and uncultivated terrestrial ecosystems

The environmental impact from the use of pesticides, understood as the effect on flora and fauna in farmland and forests is closely connected with how the individual pesticide is used and how often it is applied, the pesticide’s fate in the environment and its toxic properties. The impact on the individual plant or animal species depends on the species’ dispersal in time and space, how sensitive the species is to a given pesticide (direct effects) and how affected it is by changes in the population of other species and other indirect effects. The effect of pesticides must also be seen in relation to other factors that affect farmland fauna – particularly fertilisation, soil treatment, crop rotation and other operational measures. In addition, reduced dosages, split dosages and product mixtures affect the pesticides’ impact on the fauna directly or indirectly through their effect on the flora. The information in this section is amplified in Elmegaard (1998) and Strandberg (1998).

Exposure of field fauna through treatment with insecticides

The exposure of the fauna in cultivated fields depends in part on how the products are applied. There are three common methods in Denmark: spraying with a boom-spray, spreading on the soil as granulate and dressing of seed. Spraying is far and away the most common method. Insecticide applied with a boom-spray is normally placed in a developed crop with some degree of cover. The insecticide is deposited in small drops of water on the leaves and stems down through the crop, and the remainder lands on the ground. When cereal crops are sprayed with insecticide, 10-30% of the insecticide normally ends up on the ground. Deposition is greatest at the top of the crop, close to the nozzles, and decreases on the way down through the crop. For animals living up in the vegetation, exposure can also take place as they come and go on contaminated leaf surfaces (residual uptake) or via their food, which is either parts of plants or other arthropods.

For species that live on or in the soil, there can similarly be several exposure routes. The effect of the substance deposited on the ground depends on how strongly the substance adsorbs to organic matter and soil particles and thus also on the composition of the soil. Many insecticides adsorb strongly to clay or organic matter and therefore have low bioavailability in soil.

Direct effects on natural enemies of pests in fields treated with pesticides

Beneficial animals, meaning the natural enemies of pests, can be divided into two groups: general and specific predators. General predators live from many kinds of food and the most important species are to be found among ground beetles, rove beetles and spiders. The main species seek food on the surface of the ground and, in the case of some species of spider, by means of webs in the vegetation. The other group of beneficial animals is specific predators, which live mainly or only from a limited group of pests. In cereals, aphids are the main pests, and the specific predators are ladybirds, parasitic wasps, buzzing fly larvae and lacewing larvae.

In the case of boom spraying, direct toxic effects on field arthropods are caused mainly by insecticides, although herbicides and fungicides can also have direct toxic effects on some arthropods. The pest is often an insect living up in the crop, where it is exposed to large concentrations of the spray product. The specialised predators are also affected because they normally live in the same places as the pest. In cereals, ladybirds, buzzing flies, lacewings and parasitic wasps are therefore exposed when insecticides are applied.

The fauna on the ground can be seriously affected by insecticides in some cases, but not in others. Many of the general predators hide in cracks in the soil or under stones and similar in the daytime, when spraying is done, and are therefore hit less directly by the spray mist. The exposure occurs at night, when the animals leave their hides and wander round on treated soil and plant material (Unal, Jepson 1991).

Beneficial fauna also include bees, which, as pollinators, play an important economic role in some crops. Bees are active in the daytime and visit flowering crops to gather pollen and nectar. Daytime spraying of flowering crops with products that are dangerous for bees is therefore prohibited. This reduces the risk of direct exposure of the bees, but residual intake and intake via food cannot be avoided if the bees continue drawing on the treated crop. Some products have a repellent effect on bees, which means that the bees avoid the sprayed plants.

Effects on other non-target arthropods in farmland; general aspects

The length of time the spraying effect lasts depends on the biology of the species. Some flies and midges spend the larval stages of their life protected down in the soil or inside plants and are therefore only exposed in the short period in which they live in the open as adults. In addition, adult individuals hit by insecticides are followed by new individuals that hatch from puppae. However, for forms with larval stages on the outside of vegetation or similar and forms with lengthy adult stages, the effect can be considerable (Nielsen et al. 1996; Vickerman, Sunderland 1997).

Spiders

Spiders are often exposed to heavy effects from insecticides through direct exposure to the spray mist or through subsequent intake from the surroundings. Web-spinning spiders can gather a considerable amount of the spray product (Samu et al. 1992). Some species are very sensitive to pyrethroids (Wiles, Jepson 1992).

Springtails

The effect on springtails also depends on the form of life; species that live on the surface of the earth are more exposed to spraying with pesticides than species that live in the soil (Frampton 1988).

Effects of herbicides on the biology of the soil

By and large, herbicides have not been found to have any direct effects on the biology of the soil in laboratory tests in which field doses of herbicides have been used (Wardle 1995). However, bacteria have shown some sensitivity in around 40% of the tests. On the other hand, there are known, indirect effects on several groups of organisms. After application of a herbicide, dead organic material is added to the soil and, in the short term, stimulates microorganisms and the trophic levels that live off them. At the same time, there has been a reduction in the number of earthworms and ground beetles, which benefit from a weed-covered habitat that provides concealment and moisture. Herbicides have not been seen to have any clear effects on fungi, nematodes (but a tendency towards stimulation), mites and springtails in field tests with field-relevant doses (Wardle 1995). Lastly, the diversity of above soil level insects in a field depends on the amount of weed and therefore decreases where the weed is controlled (Wardle 1995).

There are insecticides, fungicides and herbicides that are poisonous to earthworms, but very few of them are sold in Denmark today. The carbamates are the only group of substances that is generally assumed to be poisonous to earthworms (Edwards, Bohlen 1992). Some of these substances are on the Danish market.

Leaf beetles

The knotgrass leaf beetle is a source of food for field birds. It is estimated that when a cereal is treated with the insecticide dimethoate, 7-40% of the population is exposed to a directly fatal dose of the insecticide (Kjær, Jepson 1995; Kjær et al. 1998). Practically no leaf beetles survive treatment with insecticides because the survivors eat leaves containing dimethoate (Elmegaard et al. 1998).

Direct effects of spray drift on arthropods in areas close to fields

When a field is sprayed, drift can occur to areas in the immediate vicinity of the sprayed area. The extent of this drift depends on both the spraying equipment and the climatic conditions – primarily the wind velocity (see section 4.1.6.2). A number of studies have been carried out of the effect of spray drift on the large cabbage white butterfly. These studies showed that the butterfly was affected by insecticides right up to 24 metres from the edge of the field. The effect naturally depended on the spray product used and varied between 2 and 24 metres (Davis et al. 1991, 1993, 1994; Sinha et al. 1990; de Jong, van der Nagel 1994). The long-term effect can be expected to be greater since the studies mentioned only monitored the butterflies for a short time after they had been exposed. Cilgi and Jepson (1995) thus found, for the same species, that exposed larvae continued to have excessive mortality 10 days after they had been removed from plant surfaces treated with the spray product deltamethrin.

Waysides

Waysides, like other small biotopes, have important ecological functions for plants’ and animals’ survival in the landscape. Wayside flora in Denmark comprises several hundred species and many different plant societies, ranging from totally dry to distinctly wet and from completely light and open to very shady. Studies of waysides at the end of the 1960s showed that perennial species covered about 90% of the wayside area, biennials 2%, and annuals 9%. Of these species, only the dandelion spreads seeds to annual crops. The dandelion is also favoured by the widespread mowing of waysides. Some local and county authorities still use chemical weed control in many places, over a width of about 30 cm along roads and under and around crash barriers and road signs (Bisschop-Larsen 1995). In addition, damage is often seen after herbicide treatment of fields, and in some cases the herbicide is sprayed directly out onto waysides. There is no systematic Danish registration of the effect of pesticides on wayside flora and fauna.

Direct effects on higher fauna

Direct poisoning of birds and mammals has been reported many times – also in Denmark. However, cases of individuals found dead after normal agricultural use of pesticides are rare. Many of the cases of poisoning observed and described are due to dressed seed or granulates that tempt mammals and birds. Dressed seed and granulate, lying accessible, make it possible for the animals to consume large quantities of pesticide in a short space of time. Since it can be difficult to find carrion after a case of poisoning and determine whether the death was due to poisoning, one must be cautious about rejecting such a risk because of lack of records. On the other hand, numerous field studies and laboratory tests indicate that the vast majority of pesticides used legally in Denmark do not have direct toxic effects in the concentrations in which they occur in the field. That also applies to most of the products used for seed dressing or as granulate, since products that constitute a serious threat to fauna are no longer sold on the Danish market.

Hormone-mimicking effects

Growing evidence has been found in the last few decades that man-made substances can mimic hormones or the regulation of hormones in animals and humans (Colborn et al. 1993). Such substances are often described as hormone-like substances. They can have a big effect on reproduction and the growth of organisms. Particular attention has been paid to substances that mimic or affect the sex hormones, the so-called oestrogenous and androgenous substances (Toppari et al. 1995). Manmade substances that affect the hormone systems have been known for more than 40 years and include previously and currently used pesticides and industrial chemicals. Most of the knowledge in this area concerns humans and fish, whereas little is known concerning terrestrial animals (Janssen et al. 1998). In the case of pesticides, hormone-like substances have been found among the now banned organochlorinated pesticides and organotin compounds. Alkylphenols and alkylphenolethoxylates, which have been used as coformulants in pesticide formulations, are examples of industrial chemicals that can affect the hormone systems in animal tissue (Györkös 1996; Janssen et al. 1998). Tests with these substances have been carried out primarily on rats, mice and a few bird species. However, the observed effects must be assumed also to apply to a large group of mammals and birds. According to Janssen et al. (1998), corresponding information is not available for other groups of animals. The pesticides authorised for use today are not deemed to constitute a risk to terrestrial fauna with respect to hormone-like effects. The amount of alkylphenols and alkylphenolethoxylates used as coformulants in pesticides represents less than 10% of the total use of this group of substances in Denmark. Pesticides are the only group of products for which a phase-out of these substances has been prescribed before the year 2000. The phase-out has largely been completed.

Indirect effects of pesticides on arthropods

The purpose of treating crops with herbicides, fungicides and insecticides is to remove the respective pests. Since the products generally have toxic effects beyond what is intended, plants, fungi and insects must be expected to be affected to a varying extent by spraying. This affects organisms that prey on the affected species. It has also been proven that herbivorous insects occur with much lower densities in areas treated with herbicides (Potts, Vickerman 1974; Hald et al. 1988; Hald et al. 1994; Reddersen et al. 1998) and that fungivorous insects are found in smaller numbers in plots treated with fungicides (Hald et al. 1994; Reddersen et al. 1998). In the case of predatory and parasitic insects, it can be difficult to distinguish the direct effect from the indirect effect of insecticides on their food. However, it has been found, for example, that the frequency of ground beetles with empty stomachs is higher in fields treated with insecticides than in untreated fields (Chiverton 1984).

There are thus numerous examples of pesticides affecting the density of prey, so that the predators lack food. If, on the other hand, the predators are the most sensitive to a given pesticide, the effect can be the opposite. Treatment can thereby reduce the density of predators, leading to an increased number of prey (Croft 1990). If the prey is primarily pests, there could thus be an unintended increase in the number of pests.

In several countries, including Denmark, a richer flora has been found in organically cultivated fields (Moreby et al. 1994; Hald, Reddersen 1990) and this – together with other factors that differentiate the two cultivation systems – has been followed by considerably richer insect fauna, with respect to both species and individuals, in organically cultivated fields (Reddersen 1998).

Several studies have shown a negative indirect effect on the insect fauna from herbicide spraying in addition to the effect on herbivorous insects. In particular, it is well documented that various general predators, such as ground and rove beetles, are more numerous in cereals with ground cover of wild plants (Speight, Lawton 1976; Powell et al. 1985; Chiverton, Sotherton 1991; Reddersen et al. 1998). Here, it is presumably the ground cover, which provides concealment and changes the microclimate in a favourable direction, that is the deciding factor. However, it does not appear that spiders are affected by the ground cover (Chiverton, Sotherton 1991; Reddersen et al. 1998). It has also been shown that a large number of other insect groups, including many beetles, flies and midges, are indirectly affected by herbicides via the quantity of weed. Serious and long-term indirect effects of herbicide treatment on the entire insect community in cereal fields have been observed, with greatly reduced total densities (20-85%) and also a consistently lower species diversity (Reddersen et al. 1998).

Indirect effects on mammals and birds

Indirect effects of the use of pesticides have also been found one step higher up the food chain. Common species of bird in farmland, such as partridge, pheasant, yellowhammer and skylark, breed less successfully in fields treated with pesticides than in unsprayed or organically cultivated fields, even though the substances used are not directly poisonous to birds in the dosages used (Potts 1986; Hill 1985; Petersen et al. 1995; Odderskær et al. 1997). Attention has therefore been directed towards effects on the birds’ sources of food. In the breeding season, birds and, particularly, their young eat mainly insects from all trophic levels, e.g. herbivores, fungivores and insectivores.

The relationship between source of food and the birds’ breeding success has been closely studied in Denmark in the case of the skylark (Odderskær et al. 1997). This study is one of the most detailed and statistically best designed studies with respect to clarifying the relationship between pesticides, source of food and the birds’ breeding success.

The study showed that treatment with herbicides and insecticides impairs the skylarks’ source of food, but the relationships are complex since other factors than use of pesticides affect the system. For example, the weather affects the amount of food and also affects the need for food and the time available to the parent birds to search for food, since the young cannot tolerate lying alone and unprotected in the nest in bad weather. Theoretically, one can imagine an optimum year in which the use of pesticides has no effect on skylarks because they have ample food and a catastrophic year in which pesticide treatment reduces the skylarks’ breeding success to zero. However, such extremes are likely to be very rare.

In the four years of the skylark study, the number of young leaving the nest fell on average by 38% on sprayed fields compared with fields that were not sprayed with herbicides and insecticides. The difference was due to the fact that the skylarks in treated fields had more unsuccessful attempts at breeding and largely abandoned the attempt after the insecticide treatment, whereas many pairs of skylarks continued breeding in untreated fields. For two weeks after treatment with insecticides, there was on average three times less insect food in the treated fields. Thereafter, the difference between treated and untreated fields decreased considerably. Barley fields are often treated with insecticides at the time the skylarks’ breeding activity culminates. A considerable part of the nestlings’ food consisted of ground beetles (42%), which occurred irrespective of the pesticide treatment of the fields. However, the amount of food is not necessarily the only critical variable; the quality of the food may also be of importance. An analysis of the nestlings’ faeces revealed that the content differed considerably in treated and untreated fields. The proportion of ground beetles was greatest in sprayed fields, while butterfly larvae, plant bugs, leaf-beetle larvae and several other species were more frequent in unsprayed fields, where the diet was more varied. Precisely among butterflies, plant bugs and leaf-beetles, a large proportion of the species live from wild plants. Many of the herbivorous insect species occurred in greatly reduced numbers in treated fields because their conditions were considerably impaired by both herbicides and insecticides.

In British studies of the grey partridge, it has also been concluded that the fauna living on cereals are an important part of the bird’s diet, both quantitatively and qualitatively (Potts 1986). In Britain, both partridges and pheasants have been found to have a higher survival rate in fields with unsprayed headlands (Rands 1985). The headlands are important for these gallinaceans because they often forage along the edges of fields. The edges of fields are the part of the field that is richest in species of both flora and fauna, while at the same time the crop yield is often reduced (Hald, Elmegaard 1989; Hald et al. 1994; Wilson, Aebischer 1995). The edge zone is therefore the part of the field where one gains most from not spraying.

In Denmark, the sizes of yellowhammer broods have been studied on organically and conventionally cultivated fields and found to be about 15% bigger on organically cultivated fields (Petersen et al. 1995). In this study, too, the effect is attributed to a larger, more varied and more stable food resource on organically cultivated fields, although the yellowhammer’s food was not analysed. At the same time, higher densities of yellowhammer were recorded on organically cultivated fields in the wintertime, which may be due to the more abundant flora and the consequently larger seed pool (Petersen, Nøhr 1992).

Birds are particularly interesting because a breeding bird index is prepared each year on the basis of counts all over the country. Birds are thus the only group of organisms for which we have monitoring data that can be related to pesticide consumption over time. It is known that birds are affected by a number of other agronomic variables as well and also exhibit climatically determined population fluctuations. Some species are migratory birds, which are affected in the winter period by the conditions in other countries, including the use of pesticides. To summarise, the relationship between the breeding bird index and the use of pesticides is complicated, and the use of pesticides can be regarded as one of many factors in a multifactorial complex. That means that qualitative or quantitative changes in the use of pesticides are hardly likely to be directly reflected in the year’s index but may be reflected in the trend over a number of years.

The population fluctuations of sixteen species of bird in the period 1976-1996 have been described by mathematical models as a function of a number of climatic variables, land use, treatment frequency for pesticides, size of population in the previous year and several other factors (Petersen, Jacobsen 1997). For three species – wood pigeon, sparrow and yellowhammer, the model indicated that the use of herbicides and/or insecticides had a negative impact on the population size. For these three species, a simulation of the effect of a reduced treatment frequency index (the action plan’s goal) on population size was carried out. For wood pigeon and sparrow, the simulation indicated a considerably larger population, whereas the effect on the yellowhammer population was negligible. It is not certain how these results should be interpreted, firstly because the method is based on a number of assumptions and, secondly, because it is not clear how much importance should be attached to the size of the previous year’s population. The breeding success of the skylark was thus seriously reduced by the use of pesticides in the individual year but no clear link to population development has yet been established.

Table 5.2 shows the bird species in farmland that are included in the scenario calculations in this report (see 10.3.1). The table gives the size of population and development of these species in the period from 1976 to 1996. The development has now stabilised and some farmland birds, such as the crow, have shown some progress. One of the most vulnerable and specialised farmland birds – the corn bunting – is not included because the available data are thought to be insufficient (Petersen, Jensen 1998).

Table 5.2
Characteristic farmland birds for which sufficient data are available for modelling the population development in the different scenarios. The table also shows the latest estimate of the size of each species population in Denmark and the population trend in Denmark from 1976 to 1996, using the following symbols: -: decline, +: increase, 0: unchanged (+/- 20%), 1: 20-50% change, 2: >50% change (Based on Petersen, Jensen 1998).

Repeated spraying changes the flora

Trials have shown that repeated and effective spraying of herbicides year after year reduces the number of wild plants in cultivated areas. In semi-cultivated areas, herbicides have been applied occasionally – usually in the form of products against dicotyledonous species with a view to promoting grass species and increasing the fodder value of grazing/hay production. Artificial fertiliser has also often been applied to increase fodder production on the area. This has also been done on land with section 3 status under the Nature Protection Act, which does not fully protect against adverse effects of the use of herbicides and fertilisers since the law simply "freezes" existing practice.

Spray drift

During spraying, there is spray drift to the surrounding areas, although this can be reduced with modern spraying equipment. 10-20 years ago, direct herbicide-spraying with an end nozzle on the spray boom was recommended and practised in many places. Spray drift in connection with normal use of herbicides is a particular problem for edge biotopes because of their relatively large edge area in relation to the total area. For the same reason, most of the small biotopes bordering on rotation fields are seriously affected by eutrophication, caused particularly by artificial fertiliser spread centrifugally on adjacent fields.

Herbicides and nutrients change the flora

As a result of the widespread load of herbicides and nutrients, the composition of the vegetation in these types of biotope has developed towards high-growing grass and herbaceous vegetation with low species content and completely dominated by such species as great nettle, wild chervil, common couch grass, cocksfoot grass, tall meadow oat, corn thistle, grey magwort and goosegrass. These species tolerate most herbicides relatively well and can use the ample supply of nutrients for rapid and high growth, thereby shading out most competitors. In such disturbed biotopes, one typically finds increased plant production and plant biomass per unit of area. In all, there are thus fewer species and less locally and regionally characteristic vegetation, while species with high nutritional requirements predominate.

Reestablishment of the flora is difficult

The occurrence of the most common plants in Danish fields generally became less frequent in the 20-year period from 1967-70 to 1987-89 (Andreasen et al. 1996). Weed control is practised with the precise intention of minimising some species (weeds) and favouring others (crops), and the fewer wild plants there are, the less need there is to spray the field in question. There has been a heavy decline in some species because of the intensive weed control with herbicides. In a 25-year period, from 1964 to 1989, increased use of herbicides reduced the average number of species in the seed pool from 12 to 5 per field (Jensen & Kjellsson 1995). In the same period, the total number of viable seeds in farmland halved.

With the object of investigating the relationship between the effects of herbicides on species of weed and the reduction in the number of viable seeds in farmland, a study was carried out to correlate effects at full dose of herbicide mixtures, which were frequently used in cereals in the 1970s and 1980s (effect figures from PC Plant Protection, Per Rydahl, personal communication), with the percentage decrease in the seed pool of individual species (Kjellsson & Madsen, 1998a). However, no clear relationship was found, possibly because other factors also influence the seed pool (e.g. crop rotation, soil treatment and level of fertilisation). A British study shows that the crop rotation plays a major role, affecting the size of the seed pool (Jones et al. 1997).

Spray-free buffer zones

Spray-free buffer zones are established along fields with annual crops, which are normally sprayed with herbicides and other pesticides as described in section 5.1. The wild flora generally reacts strongly and immediately to spray-free conditions because there is plenty of seed in the seed pool to ensure a high plant density. Many wild plants are characterised by a flexible and largely individual growth potential. One could in particular expect considerable quantitative changes and, only secondarily, a number of qualitative changes in the flora. For example, in buffer zones in winter wheat, the total biomass of wild plants was seen to increase 10 to 40 times within a year (Reddersen et al. 1998). The increased growth and seed production would also significantly increase the seed pool after 2-3 years (Hald et al. 1994; Kjellsson & Rasmussen 1995). Buffer zones in cereal crops would have a beneficial effect on common dicotyledonous plant species such as shepherd’s purse, corn pansy, forget-me-not, speedwell, chickweed, white goosefoot and dead nettle, together with a large number of less common species. At the same time, some grasses – particularly annual meadow grass – would decline somewhat. In the short term, one could thus primarily expect an increased total biomass of wild plants caused by the increase of a rather limited number of common species, while an improvement for the less common species would generally require more permanent buffer zones. Hald et al. (1994) thus did not observe any clear increase in the number of species in permanent buffer zones over 5-6 years.

Permanent unsprayed, fertiliser-free headlands

Like buffer zones, unsprayed headlands could help to protect well-preserved vegetation in small biotopes, where such vegetation still exists. In most places, however, the vegetation in small biotopes has been seriously affected by both herbicides and fertilisers in the last few decades. To ensure recolonisation, which normally takes place very slowly, it would be necessary to establish permanent spray-free and fertiliser-free edge zones.

Effect on seed production

It has been shown that sub-lethal doses of herbicides lead to a reduction in plants’ seed production. The reduction is related to the dose, and this is probably directly related to a smaller biomass production (Rasmussen 1993; Andersson 1994). For example, a sub-lethal dose (1/2 the normal dose) of isoproturon resulted in a 50 per cent reduction in seed production in common pennycress (Hald 1993). In a Danish study it was shown that seed production in unsprayed field plots was 6-14 times higher than the seed production in sprayed plots (Kjellsson, Rasmussen 1995). At the same time, spraying with herbicides (dichlorprop + 2,4-D/MCPA in normal dosage) resulted in a smaller proportion of the surviving plants being able to propagate. It has been shown that some herbicides (tribenuron-methyl and, to a lesser extent, MCPA) can result in a smaller seed size in some species, such as black bindweed, goosegrass and common pennycress (Andersson 1994). With this knowledge and different scenarios for herbicide use (0 and 100%), it would be possible to model and estimate the effects on the seed pool in farmland and the consequences in the form of changes in the frequency and composition of the vegetation (Kjellsson, Rasmussen 1995; Madsen et al. 1996, 1997, 1999).

The wild flora in hedgerows and small biotopes

Herbicides are normally not used in hedgerows and small biotopes, so any impact on these areas is due to unintended effects from agricultural use of herbicides, e.g. through spray drift (see section 4.1.6.2). During a 5-year period, the flora in Danish field hedges showed a slight tendency to contain more species along unsprayed edges of fields than along sprayed ones (Hald et al. 1994). A 50% standard trial treatment with the herbicide fluroxypyr of a fallow field on sandy soil in the Netherlands reduced the content of species (Kleijn, Snoeijing 1997). Effects on survival at lower dosages (5 and 10 %) were only found for a few species in some years. In another Dutch study, de Snoo (1997) found, in a three-year trial, increased species diversity in unsprayed field edges in sugar beet, potatoes and winter wheat, primarily as a consequence of an increase in the number of dicotyledonous plants.

Effect on the flora through spray drift

There exists a set of standardised values from Germany for spray drift. The values are based on 16 field trials in the period from 1989 to 1992, where, on the basis of a 95% percentile, a so-called "realistic worst case" was established (Ganzelmeier et al. 1995). On the basis of these values and known effect thresholds (PC Plant Protection), a qualitative analysis can be carried out of the effects on plant growth. The method used by Ganzelmeier is not necessarily ecologically relevant. The deposition was measured during spraying with a single spray plume, and the spray product was collected on flat targets laid on the ground. Such data are relevant for estimating the effect on plants that have not yet germinated and for "flat" areas like ponds and lakes. However, plants that have germinated and are established "catch" more of the spray product. Both Davis et al. (1993) and Bui et al. (1998) found that different "targets" had different "catch efficiencies". "Targets" that rise above the ground and have a complex structure catch more spray product than objects lying flat on the ground. Finally, Nordby and Skuterud (1975) found that the dose of spray product needed to trigger a given effect is smaller for plants in the drift zone than for plants directly under the sprayer. An American study showed that sweet cherry is damaged by herbicide drift (chlorsulfuron) at doses down to 1/100 the normal field dose (Al-Khatib et al. 1992). At doses 1/3 to 1/10 the normal dose, several herbicides (2,4-D, glyphosate) caused significant damage. A number of studies (Marrs et al. 1989, 1993; Davis et al. 1993, 1994) of the effect of spray-product drift on plants showed that plants were affected up to 50 m from the sprayed area. However, the majority of the plants were only affected over a distance of 0 to 5 m from the field.

Effects on forest-floor flora

Actual studies of the effect of pesticides on forest-floor flora are rare, so the assessment has been based on knowledge concerning the mode of action of herbicides and knowledge concerning the ecology of forest-floor flora. It is only the effect of herbicides that has been considered because, according to Elmegaard et al. (1996), there are not stated to be any known effects from other groups of pesticides on forest-floor flora.

Østergaard et al. (1998) state that glyphosate is applied once before clear cutting of deciduous trees and conifers and once or twice in the first few years thereafter. This application practice is thought to have a radical effect on the forest-floor plants, so that all individuals of the flora associated with the type of forest in question may be eradicated (F. Rune, Danish Forest and Landscape Research Institute, personal communication). There will still be a seed pool, but it will also be seriously reduced, partly because of the repeated spraying and partly because of changes in the forest climate for a time when renewal takes place by means of clear cutting. Plants from seeds that germinate in the first period after clear cutting, possibly provoked into germinating by the increased amount of light, will often die, either as a consequence of spraying or because of drought, frost-nipping or scorching by the sun. In addition, forest-floor plants have a short-lived seed pool, so these species have no possibility of surviving an unfavourable period during the seed stage (Graae 1999). Even if the treatment is carried out over a limited number of years, it will have a serious impact on the forest-floor flora. Seen over a rotation period, the relatively small treatment frequency index thus has a big impact. In Christmas tree and ornamental greenery cultures, where herbicides are used during the entire lifetime of the culture, there is virtually no forest-floor flora. So massive are the effects that there is no flora at all – which is, of course, the whole idea of using herbicides in these cultures. When Christmas trees and ornamental greenery are produced without herbicides, a flora develops that is dependent on the way the soil is treated, and the method of renewal can be more or less authentic (M. Strandberg, DMU, personal communication).

The rate of recolonisation by forest-floor flora species is very slow (0 - 1m/year) (Brunet & von Oheimb 1998), so it is only in areas in the immediate vicinity of the forest with elements of the original flora that one can expect partial recolonisation by the species that have disappeared. Indeed, Brunet and von Oheimb (1998) also recommend that the slow recolonisation rate of forest-floor flora be taken into account in forestry planning. Inghe and Tamm (1985) have shown in Swedish forests that individuals of blue anemone are more than 40 years old, and it is not unlikely that many species of forest-floor flora can reach a higher age than trees. Danish studies have shown that forests with long continuity (>200 years) have a better developed forest-floor flora than younger forests (Graae 1999), so for Danish forests, too, it is reasonable to expect impacts on the forest-floor flora if herbicides are used in connection with clear cutting and reestablishment of cultures. According to Graae (1999), species propagated mainly by clonal growth will recolonise very slowly, while those dispersed by animals and the wind recolonise more quickly. Mechanical soil treatment instead of the use of herbicides presumably has similar effects on the forest-floor flora, but this has not yet been studied.

Impact on the flora of nature areas through atmospheric transport of pesticides

Herbicides are not normally used on nature areas, but many herbicides are transported over long distances and small amounts of them are thereby deposited on nature areas. No information has yet been found on measurements of direct impacts from pesticide drift on the flora in nature areas apart from boundary areas such as hedgerows. It is therefore necessary to use model calculations. In the Netherlands, the average combined herbicide consumption is 1.35 dose equivalents per year, 5.5% of which evaporates. On this basis and import/export considerations concerning atmospheric transport, it was calculated in a Dutch model study (Klepper et al. 1998) that the Dutch nature areas received an average of 0.02 dose equivalents per year. By means of a dose-response model for the potentially affected natural vegetation, this deposition was used to predict how large a percentage of the species were affected beyond NOEC (No Observed Effect Concentration). The result was that 2% (median value) of the species were affected beyond their NOEC. The percentage of affected species was highest in agricultural areas and in areas with fruit and berry growing. There are no similar calculations for Danish conditions, but the use of herbicides with a treatment frequency index of 1.65 in 1997 (DEPA 1998a) is comparable to the Dutch consumption, since the frequency index corresponds to the Dutch dose equivalent. It cannot be concluded directly from the Dutch calculations that around 2% of the species in Danish nature areas are unacceptably affected by herbicides. The level of impact depends, in part, on how the nature areas are situated in relation to areas in which herbicides are used and on the magnitude of the emission, the wind conditions and the sensitivity of the local plant community. In the Dutch calculations, the biggest uncertainties were reportedly due to the emission calculation and the effect models (Klepper et al. 1998). A diffuse dispersal of pesticides, e.g. via the atmosphere with precipitation, must be regarded as having less effect on the composition of the flora in nature areas than increased supply of nutrition and changed nature management.

The effect of pesticides in rainwater

The occurrence of pesticides in rainwater in Denmark is discussed in section 4.5. Herbicides have also been detected in rainwater in countries in Scandinavia and Northern Europe (Kirknel, Felding 1995ab), but no direct effects on flora as a consequence of this have yet been found (Felding 1998b). It is known from American studies that deposition of atmospheric herbicide residues (sulphonylurea) can produce symptoms of damage in some crops, such as peas and beans (Felsot et al. 1996). It has similarly been shown that even small doses of chlorsulfuron (1/100th to 1/1000th part of normal dose) greatly reduce the plant biomass and seed production in persecaria (Fletcher et al. 1996). In a Danish project on the occurrence of pesticides in precipitation and effects on plants and plant communities, the effects of mechlorprop have been investigated in concentrations corresponding to those found in rainwater. The study covers sensitive species – particularly the crucifers – but no effects have been found (Solveig Mathiassen, DIAS, personal communication). One can assess possible effects by combining data for the deposition of pesticides with effect data.

5.2.1 Conclusions

Together with crop rotation and other cultivation measures, the use of herbicides has considerably reduced the flora in farmland under cultivation. Accidental spray drift can have adverse effects on plants in hedgerows and small biotopes. In particular circumstances, herbicides in rainwater can presumably cause damage to plants outside the cultivated area.

Pesticides can be transported to watercourses via the atmosphere, through wind drift from field spraying or long-distance transport, via groundwater, drain water and surface run-off, and through unlawful spraying in or near watercourses within the 2-metre buffer zones. From the watercourses, the pesticides can be led to lakes or coastal waters. These can also receive pesticides by the same transport paths as watercourses. All transport of pesticides to fresh water in Denmark is undesirable. Data are available on the occurrence of pesticides in watercourses and ponds, as described in sections 4.2 and 4.3, but there have not yet been any systematic studies of the occurrence in Danish lakes and coastal waters. Studies in lakes have been initiated in connection with Aquatic Environment Plan II.

In this section we describe the impacts of pesticides in fresh water. Since there are only a few Danish studies, the description is based primarily on foreign studies, with the main emphasis on field studies (Friberg 1998). In order to assess the effects of current practice for use of pesticides, concentration levels given in Mogensen and Spliid (1997) are combined with results from the literature. Denmark has distance requirements for many pesticides, which must not be applied closer than 10 or 20 metres to lakes and watercourses. Table 5.3 shows the active ingredients to which such distance requirements currently apply. Some of these ingredients affect organisms in the aquatic environment at lower concentrations than 1 microgramme per litre and a few at lower concentrations than 0.1 microgramme per litre.

Table 5.3 Look here!
Pesticides subject to special distance requirements. The distance requirement is generally 2 m to watercourses and lakes. The table shows the pesticide, its active ingredients, the type (F = fungicide; H =herbicide; I = insecticide; P = growth regulator). Also shown is the lowest concentration for individual substances that cause 50% mortality or inhibition of activity in laboratory tests with aquatic organisms (fish, crustaceans and algae) (m g/L = microgramme per litre).

The effect of pesticides depends on whether the recipient is a watercourse or stagnant water. The effect of pesticides on aquatic organisms generally depends on the retention time and thus on the exposure time. In watercourses, the retention time depends on the average rate of flow and the presence of areas with a low rate of flow. In a watercourse to which a herbicide was added, a stretch with a high rate of flow directly downstream from the treated area was analysed and compared with a more slow-flowing stretch 225 m downstream (Thomson et al. 1995). Although the concentration was briefly higher in the upstream stretch in connection with the spraying, the retention time of the substance in a concentration of more than 1 microgramme per litre was twice as long in the stretch of water 225 m downstream. In watercourses, the concentration of pesticides also decreases at a given distance downstream of the source, depending on dilution and deposition/metabolism. In stagnant water, the effect of the pesticides depends particularly on the volume of water and the rate of water replacement.

Small and slow-flowing watercourses and small ponds are therefore most exposed to any pesticides. Moreover, these systems are closely linked to the surroundings and are therefore more likely to be affected by any use of pesticides in the surrounding areas.

Biologically, watercourses differ from stagnant waters by being open systems. There is permanent drift of organisms with the current, which means that stretches of watercourses are generally recolonised quickly. That means that watercourses are generally very resilient and therefore quickly return to normal once a chemical impact has ended.

Effects on primary producers

There are many documented instances of effects of herbicides on primary producers in freshwater ecosystems, whereas insecticides seldom seem to have any direct effect on plants. The effect of herbicides is usually that they reduce productivity, whereas, in the short term, the biomass is not affected. There is a clear tendency for the adverse effects to end very quickly when the exposure is over. Besides the direct effects of herbicides on primary producers in watercourses, indirect effects have been found in a number of studies. For instance, photosynthesis in an algae community increased at a concentration of more than 4 microgrammes per litre of the insecticide lindane owing to a reduction in the number of grazing invertebrates (Pearson, Crossland 1996).

With the concentrations found for the herbicide atrazine, it is unlikely that algae and macrophytes will be affected. Generally speaking, the concentration has to be much higher than 50 microgrammes per litre before the ecosystems are affected (Solomon et al. 1996). In the case of the herbicide hexazinone, the measured concentrations are also so low that no serious effect on the primary producers is expected. However, up to 43 microgrammes per litre have been found in a drain water sample, which could affect aquatic plants locally if there were little dilution in the recipient watercourse. EC₅₀ (4 hours) for hexazinone has been found to be 3.6 microgrammes per litre for aquatic plants in watercourses, and a concentration of 145-432 microgrammes per litre reduced the productivity of aquatic plants by 80% (Schneider et al. 1995).

Effects of insecticides on fauna

A large number of studies show that pesticides can affect the fauna in freshwater ecosystems, both directly and indirectly. Generally speaking, insecticides, in particular, have an adverse effect on the fauna, whereas direct effects from herbicides are often seen only at relatively high concentrations. Indirect effects are seen particularly in the highest trophic levels (fish, amphibians and birds) owing to reduced quantities of prey or in the form of changes in the function of the entire ecosystem. For example, Wallace et al. (1991) found that the decomposition of leaves into fine organic matter fell considerably in a forest watercourse after treatment with the insecticide methoxychlor owing to elimination of the insects that comminute the leaves. Methoxychlor may no longer be used in Denmark. If the possibility of recolonisation of a watercourse is poor, e.g. owing to barriers or the geographical location, it can take a long time (years) for the watercourse’s biological conditions to be reestablished, even if the supply of pesticide ceases. In the case of stagnant freshwater systems, the effects of pesticides seem to disappear within a similar period of time. These ecosystems are generally affected by both fertilisers and pesticides.

Effects of herbicides on fauna

In the case of the herbicide atrazine, the concentrations found in watercourses are generally so low that no effect on the fauna is likely. For example, in a study by Grande et al. (1994), the mortality among trout fry, which are more sensitive than adult fish, only increased at atrazine concentrations of more than 50 microgrammes per litre. However, Lampert et al. (1989) showed that concentrations right down to 0.1 microgramme per litre could affect daphnia in stagnant water, even though, in a single-species test, EC₅₀ was 2 microgrammes per litre. The study shows that the sensitivity at community level is far greater because both direct and indirect effects play a part. Atrazine concentrations of more than 0.1 microgramme per litre have been found in a pond near Køge, but there are very few other measurements. As in the case of atrazine, the concentrations of hexazine that have been found are so small that no serious effect on the fauna is likely.

The fungicide propiconazole has been found in concentrations up to 0.8 microgramme per litre in a watercourse and 0.1 microgramme per litre in a pond. These concentrations are below the concentration found to have an effect on zooplankton, invertebrates and fish.

The insecticide dimethoate has been found to affect (increase) activity in watercourse invertebrates at concentrations of more than 1 microgramme per litre and to reduce the density of some species (Bækken, Aanes 1994). In the case of zooplankton, an LC₅₀-value of around 20 microgramme per litre has been found in mesocosm experiments (Hessen et al. 1994). The concentrations in Danish watercourses and lakes lie below these values but not significantly below the concentrations that produce effects.

The pyrethroid insecticides are generally very toxic to largely all freshwater fauna, i.e. to zooplankton, invertebrates, crustaceans, fish and amphibians, in very small concentrations – normally less than 1 microgramme per litre. Besides that, the substances accumulate in the organisms (e.g. Andersen 1982), but are eliminated when the exposure ends. The pyrethroid fenvalerate has been found in ponds in a concentration of 0.12 microgramme per litre. Anderson (1982) observed behavioural changes and mortality in invertebrates – particularly amphipods – at fenvalerate concentrations from 0.022 microgramme per litre.

Effects on amphibians

There are only a few studies of the effect of pesticides on amphibians. In a Danish study, the insecticide esfenvalerate caused paralysis in embryos of the fire-bellied toad and the xenopus at a concentration of 1 microgramme per litre. At 5 microgrammes per litre, 80% of the embryos of the xenopus had deformed bodies, oedema and deformities of the spine, brain and intestine (Larsen, Sørensen in prep). Esfenvalerate was detected in a stream, Lillebæk, on Funen in the period 1994-1996 in a concentration of 0.2 microgramme per litre. (Funen County 1997). In that period, the recommended field dosage of esfenvalerate was 25 grammes per hectare. In a pond test in northern USA, esfenvalerate was found to have an adverse effect on daphnia and copepods at a concentration of 0.01 microgramme per litre (Lozano et al. 1992).

Unlawful use of pesticides

With our present level of knowledge, it is not possible to determine how large a part of the measured concentrations of pesticides in fresh water are due to unlawful use.

Watercourses must generally be deemed to be very sensitive to accidents with pesticides at washing sites and similar because, in many cases, the water is led directly to the nearest watercourse. Danish watercourses are also generally small, i.e. more than 80% of them are less than 2 metres wide, and they have low rates of flow. Therefore, even a short spillage of pesticides could have a major, acute effect on the ecological conditions in them.

Effects of pesticides in rainwater

On the basis of rainwater data given in Mogensen and Spliid (1997), the pesticide content must be assumed to be generally so low that it will have no effect on the biological conditions after further dilution in the fresh water system (irrespective of size). However, that does not exclude the possibility of drift from neighbouring fields resulting in local increases in the concentration of pesticides to a level that can be harmful to the freshwater environment. All the same, this problem is probably limited provided the distance requirements for spraying are observed.

Other environmental factors and pesticides

The effects of a given pesticide depend on the other ecological factors in the freshwater environment. For example, Caux and Kent (1995) found that a green alga Selenastrum capricornutum was affected differently by atrazine, depending on the chemical composition of the water. The role played by this has not been taken into account in this report. Furthermore, pesticides in watercourses often occur in pulses, especially with unlawful use. It is during such an episode that the concentration of pesticides must be measured for the effects to be assessed. This is seldom possible because the water containing the pesticide passes the sampling site within a very short space of time, so the chance of extracting a sample of the polluted water is extremely small. The measured concentrations are therefore probably often lower than the maximum that has occurred in the system. Another factor that probably has a significant influence on the effect of pesticides in watercourses is the physical conditions. Most Danish watercourses have been physically altered to ensure drainage of surrounding fields, whereby they have been made unnaturally wide and physically uniform. Both the degradation time and the effect on the ecological conditions presumably depend on the physical heterogeneity, so – all else being equal – physically uniform watercourses will be affected most by any supply of pesticides.

5.3.1 Conclusions

There are very few Danish studies of the occurrence and effects of pesticides in Danish watercourses and ponds and, as yet, no systematic studies in lakes and coastal waters. The assessment of the effects of pesticides has therefore been based primarily on foreign studies, and importance has been attached to field studies. It has not been possible, either, to find any studies of the effects of all the pesticides detected in Danish freshwater ecosystems. Since there are often differences in the species composition and the structure and function of the ecosystems in Denmark, compared with the foreign studies, the sensitivity to a given pesticide should be treated with caution. Owing to the relatively low water temperatures in Denmark, compared with more southerly countries, the degradation time for pesticides is shorter in Denmark, so the exposure time and thus the risk of effects are greater. With our present level of knowledge it is therefore difficult to judge how the present use of pesticides affects Denmark’s freshwater systems. However, several measurements indicate that pyrethroids and thiophosphate insecticides have been detected in concentrations at the level that has effects according to the existing literature. In particular, the available concentration levels indicate that it is insecticides – and especially the pyrethroids – that can have an adverse effect. In addition, by reason of their persistence, the pyrethroids can occur in freshwater ecosystems for a long period, during which they can be absorbed by, for example, invertebrates that live from dead organic material.

In connection with authorisation of pesticides, tests are carried out to determine their toxic effect on individual species of algae, daphnia, fish and other organisms. Mesocosm analyses that simulate entire ecosystems are also performed. However, these analyses do not provide insight into the many natural factors that interact in nature and they do not cover the combination of the many different pesticides detected in the aquatic environment.

It is therefore likely that the freshwater environment is affected by the present use of pesticides, but it is not yet possible to determine the magnitude of the impact owing to lack of data and knowledge. The county authorities, in their regulatory capacity, have provisionally estimated that around 2% of the unfulfilled targets in approx. 11,000 km of watercourses are due to toxic substances, including pesticides (Windolf 1997). However, this figure is based on a subjective evaluation and will also vary with the region, the sampling method and the frequency.

The main impacts occur in connection with the application of pesticides, with organisms being hit directly, and with indirect effects occurring as a consequence of the effect on food chains. Here, plants play a key role as the first link in the food chains. A Danish study has shown that the number and frequency of plant species in field studies have halved in the last 20-25 years. From an agricultural point of view, this has been a desirable development, but it has adverse consequences for the natural species. The main reason for the decline is the use of herbicides, but changes in cultivation practice, including the use of fertilisers, have also had a major effect.

During spraying, there is spray drift to the surrounding areas. However, hedgerows, dikes, dry stone walls and other small biotopes are of such small width that they should in practice be included in the area that is affected by spray products. Spray drift can affect both terrestrial and aquatic ecosystems. In the case of the aquatic environment, any effect from pesticides is undesirable, including changes of flora and fauna in coastal waters, lakes, ponds and watercourses. Among the aquatic ecosystems, it is particularly ponds, watercourses and lakes near fields that could potentially be affected.

The freshwater environment is in all probability affected by the present use of pesticides, but it is not possible on the basis of the existing data to quantify the magnitude of the impact nationally. On the basis of information from the county authorities, it is provisionally estimated that around 2% of the unfulfilled targets of about 11,000 km watercourses can be due to toxic substances, including pesticides. In particular, the concentration levels measured indicate that it is insecticides – and especially pyrethroids – that can have an adverse effect. Owing to their persistence, pyrethroids can also occur in the freshwater ecosystems for a long period of time. There are also documented cases of effects of herbicides on algae and other primary producers. Several measurements indicate concentrations of pyrethroids and certain thiophosphate insecticides that are close to the level that has effects according to the existing literature. For some pesticides, this level is lower than the limit value of 0.1 microgramme per litre for drinking water.

In both cultivated areas and the adjacent biotopes, the use of pesticides involves a risk of reduced populations of flora and fauna, changed biodiversity, changes in the cultivation medium and natural pest regulation, and of food-chain effects and other indirect effects. Seen overall, it is not the individual field and its possible loss of wild flora that are the problem, but rather the total, countrywide effects on characteristic farmland flora and the associated fauna.

In forestry, very little use is made of pesticides, whereas, in Christmas tree and ornamental greenery cultures, the same quantities are used as in farming. The treatment frequency index in nurseries and market gardens is high. There is a lack of specific studies of the effect of herbicides on forest-floor plant species, but there is no doubt that even the limited use made of pesticides in forestry has adverse effects. Many woodland plant species have a very slow recolonisation rate of less than 1 metre per year, which makes them particularly sensitive to herbicides, even though these are only applied in connection with clear cutting and afforestation.

The sub-committee’s recommendations concerning the impacts of pesticides in the environment

For the scenarios in which pesticides are used, there are no systematic studies of how pesticides in large, connected areas affect wild flora and the associated fauna in hedgerows, waysides and other small biotopes and in neighbouring nature areas. The effects on the flora caused by precipitation of long-distance transported herbicides in Denmark is not known. Foreign studies show that effects probably exist, but further investigations of both the atmospheric transport and the effects are needed for any closer determination. There is also a need to assess the effect of pesticides on aquatic organisms in relation to the actual finds in watercourses and surface water. The sub-committee recommends that the necessary knowledge be built up and that time series be established to document any effects on the terrestrial and aquatic ecosystems seen in relation to the proven occurrence of pesticides in the environment.

The sub-committee recommends more consistent and systematic use than hitherto of permanent unsprayed headlands and protection borders as buffer zones to help protect watercourses, lakes and ponds, together with well-preserved vegetation in small biotopes and nature areas, where such still exist. In this connection, it must be ensured that linked dispersal corridors are established. Where the vegetation of small terrestrial biotopes has been seriously affected by the last few decades’ intensive use of both herbicides and fertilisers, recolonisation will normally take a very long time. Here, it will be necessary to establish both spray-free and fertiliser-free boundary zones where restoration of the vegetation and the associated fauna is desired. The sub-committee also recommends nature restoration by sowing of wild plants and the introduction of species of fauna with a low dispersal potential.

6. Exposure of humans and health effects

6.1 Exposure of, and effects on, agricultural workers

Agriculture and health and safety

Primary agriculture comprises farms, market gardens, poultry farms, fur farms and nurseries. The total number of persons employed is approx. 96,000, corresponding to just under 4% of total employment. Of that figure, about 12,000 are employed within market gardening. In 1997 there were 60,900 farms with an average size of 44 ha. Of these, about 400 farms have five or more employees (Danmarks Statistik 1998). As at 1 January 1999, these farms are required to have a safety representative. The average age of farmers is 52 years.

The treatment frequency index for pesticides in market gardening and fruit growing is considerably higher than in farming. The index is thus typically 20-25 in the production of apples, 11 in strawberries and 4-12 in vegetables. In nurseries, the index is 4-14. There are no figures for the treatment frequency index in greenhouses, but it is presumably higher than the outdoor index. An exception is greenhouse production of cucumbers and tomatoes, where biological control is used.

Health and safety factors in weed control

Table 6.1 shows factors that are of importance to health and safety with and without pesticides. It is mentioned in the table that vegetables need manual weeding as well as mechanical weed control. For example, carrots have to be weeded manually once per season and onions twice per season. In large production units, a trolley is used. On this, 10-15 people lie on their stomachs with their heads supported. The trolley is pulled along at a speed of about 600 metres per hour over the rows, which are weeded by hand.

Table 6.1
Factors of importance to health and safety with and without pesticides (Based on Bjørn, Rothmann 1998).

General ergonomic problems

Driving a tractor can seriously load the driver’s back because he often has to sit with his back turned in order to keep an eye on the work. This applies not only during ploughing, harrowing and other forms of weed control, but also during sowing and spraying. In connection with sowing there is a risk of back injury because sacks of seed weigh 50 kg. In addition manual collection of fieldstones involves a risk of physical injury, both from handling heavy stones and from use of a stone fork for smaller stones. Stable work, including handling of farm animals, is often hard, physical work. It seems to be generally accepted by farm workers that their work involves a serious risk of physical wear and of injury. The most common health complaints are "back problems", "bad foot" and "squeezed fingers". On the other hand, it seems that young farm workers are no longer willing to expose themselves to heavy loads or to use methods that are regarded as physically wearing.

Particular problems with pesticides

The risk in connection with the use of pesticides can be considerably reduced by using suitable safety equipment and by following the prescribed safety rules. Even so, some users are not happy about using pesticides because of doubts concerning the efficacy of gloves, breathing equipment and other safety equipment (Bjørn, Rothmann 1998). Since it is not possible in practice completely to avoid contact with the products, some users are worried about possible chronic effects. Others put ensuring the harvest ahead of any long-term effects, so the working environment does not have the highest priority.

6.1.2 Accidents and injuries

Reported occupational injuries

The Danish Working Environment Authority (WEA) has analysed the statistical data for the period 1993-1997 concerning reported occupational injuries in farming (Bjørn, Rothmann 1998), using employment figures from farming as at 1 January 1994. In the three largest sectors, the employment figures were as follows:

1. Dairy farming 29,564 persons
2. Arable farming 25,498 persons
3. Mixed farming 21,284 persons

About 75% of the people working in agriculture were employed in these three sectors, and it is also these three that have the biggest number of reported industrial accidents.

In the period 1993-1995 26 fatal accidents were reported, 20 of which occurred in the said three sectors. Most of the serious accidents reported – resulting, for example, in amputation, broken bones or injury to extensive parts of the body, also occurred in these three sectors, namely, 256 reports out of 395. For the whole of agriculture, occupational injuries were reported for 1,318 persons.

The largest number of "very serious accidents" (331 out of 395) was reported for persons of 45 or under. The largest number of fatalities (21 out of 26) was reported for persons over 45. It should be noted that even though about 50% of the employees were over 50, it is the younger age groups that had the highest report incidence (8-12 reports per 1000 employees). The average incidence for reported industrial accidents in the whole of agriculture was 4 reports per 1000 employees. The age groups 18-24 years, 25-29 years and 30-34 years accounted for 54% of all reported industrial accidents. The 18-24 year age group accounted for 27% of reported, very serious industrial accidents. Most of the industrial accidents occurred in August-September.

In the case of children and young people under the age of 18, most reports related to the 16-17 year age group. 84% of children and young people employed in agriculture were less than 18 years of age. For this group, most of the industrial accidents were reported within the first six months of employment. Most of the reports concerning children and young people were received in June-July. Farming has more fatalities than any other sector of industry.

Type of accident event

About one third of the types of accident events are classified as "lost control of machines, aids and systems", but accidents in connection with work with animals (aggressive animal/animal out of control) and accidents involving falls also together account for one third. In the period 1993-1997 a total of 2,429 industrial accidents and 979 cases of work-related diseases were reported. Only the material for the period 1993-1995 has been thoroughly analysed.

Tractor accidents

In the period 1993-1995 there were 26 fatal accidents. In 12 of them the cause was "lost control of machines, aids and systems", and in six, a tractor was involved. Of four fatalities due to falls, two were falls from a tractor. One of two fatal traffic accidents was caused by a reversing tractor. In the period in question, 72 industrial accidents with tractors were reported. These accidents are often very serious. There are relatively many fatal accidents, namely 9 out of 72 in which tractors are involved.

Accidents resulting in musculoskeletal injury

Under the classification "lost control of own movement" (musculoskeletal injury), 79 industrial accidents were reported in the period 1993-95, six of which were serious (primarily broken bones). These accidents frequently occur during lifting, pushing and/or pulling of objects and animals.

Occupational diseases

Most of the occupational diseases reported in the period 1993-1995 were within the diagnostic category "musculoskeletal diseases". They were followed by hearing damage and skin diseases and then by pulmonary diseases, both allergic and non-allergic. One case of cancer was reported in the period. It has not been possible to find more than one named pesticide exposure, namely methylparathion, listed under "other diseases". There is no mention of the specific diseases that resulted from this exposure. A few other pesticide exposures may be listed in the category "chemical effect without specification".

The "Ringkøbing study"

In view of the large number of accidents that occur in farming, Ringkøbing County’s Occupational Medicine Clinic and Herning Central Hospital, in cooperation with the Agricultural Advisory Service, have analysed possible causes of accidents in farming. The project was divided into three phases (quoted by Bjørn, Rothmann 1998).

First phase

The first phase consisted in recording all serious farming accidents in Ringkøbing County. This was done in 1992. 257 serious occupational accidents were found, four of which were fatal. There are around 8,000 farms in the county.

Second phase

In the second phase, about 400 farms with about 1,600 farmers, farm workers, spouses and children were chosen at random from three Danish Agricultural Advisory Centres in Herning and Holstebro. For one year (1993-1994), the farmers were asked to record once a week both small and large accident events occurring in connection with farm work. At the same time, the hours spent on the different types of work, such as field work, stable work, repair work, etc., were recorded. During the year, 389 accidents resulting in injury were recorded. Of these, 28% required medical treatment. After adjustment for work time used, the work-related accidents hit all age groups equally. 62% of all independent farmers had one accident resulting in injury per year. The corresponding figure for farm workers was 22%. 45% of the accidents occurred in connection with direct contact with animals. When fall accidents in stables and work-related accidents during use of stable machines are included, accidents in stables accounted for 51% of all accidents.

Animals require many hours of work. Taking this into account, the frequency of accidents with animals and machines is almost the same. With this time weighting, repair and maintenance involve an almost seven times higher risk of a work-related accident than ordinary farm work.

Most work-related accidents occur in the autumn, when most hours are also spent on work on the farm. With more working hours, the number of work-related accidents increases, but adjusting for the increased risk time, one finds no increase in frequency due to long working hours.

Third phase

In the third phase, a safety inspection of the farms was carried out. At the same time, a team of 10-15 instructors held a safety course at which work habits and methods of work were discussed. Before this preventive intervention, the farms in the test group had 29.2 accidents per 100,000 working hours. After the intervention, this figure fell by just under 40% to 17.5 accidents per 100,000 working hours. In the control group, 21.3 accidents per 100,000 working hours were recorded at the start of the same period, and 20.0 accidents per 100,000 working hour at the end of it. More serious work-related accidents were also reduced by about 40% in the test group. In the analysis, a relationship was found between stress and work-related accidents, with the persons indicating the most stress symptoms also being those with the most accidents. Those participating in the project were asked about the time they spent on field spraying. The average was 39.5 hours per farm per year. For the individual farm, the time spent on spraying per year varied from 0 to 1,133 hours (Bjørn, Rotmann 1998).

The organic sector’s status in South Jutland

The working environment at organic farms was included in a special questionnaire-based survey in South Jutland County among farmers from both organic and conventional farms and among agricultural advisers. The survey, which covered 90 persons, is reported in Bjørn and Rothmann (1998). In all, 48 responses were received.

36 responses came in on health and safety. Of these, 24 farmers stated that the organic form of production had not involved more manual work. 12 farmers thought it had given rise to more manual work in the form of feeding, lifting beets, moving cows and spreading straw. One farmer mentioned osteoarthritis, which figures as the only stated work-related injury.

In reply to the question of whether the restructuring had caused problems with dust, 41 persons replied "no", while two farmers said "yes" and mentioned spreading of straw and hoeing of potatoes. In reply to the question of whether they had experienced a beneficial effect from no longer using spray products, 26 farmers replied that they had, while 8 farmers replied that they had not experienced any beneficial effect. The main reasons given for switching to organic farming were a desire for new challenges and suspicions about pesticides.

6.1.3 Work-related factors with pesticides and mechanical methods

The sprayer operator’s exposure to pesticides

There is extensive knowledge concerning the exposure of sprayer operators in different spraying scenarios (EUROPOEM 1997). This knowledge has been put on the Internet (DIAS 1998). Figures 6.1 and 6.2 show scenarios for an average Danish farming scenario with use of a tractor-drawn boom sprayer, where the sprayer operator fills the tank and sprays, using either a spray fluid (figure 6.1) or a powder pesticide product (WP, wettable powder) (DIAS 1998). It is assumed in the scenarios that the sprayer operator wears protective clothing and that the gloves and the clothing retain 50%. A daily work time of 6 hours is assumed, during which 20 ha are sprayed. There is potential exposure of the skin, i.e. the amount of pesticide lying on the skin, since the amount absorbed by the organism depends on the ability of the individual pesticide to penetrate the skin. It will be seen from the figures that 85-99% of the exposure occurs during filling of the tank, even though this work only accounts for a small part of the total working time spent on spraying.

Figure 6.1
The tractor operator’s exposure to liquid products via inhalation, hands and body in different work operations (filling of tank and the actual spraying). The total exposure for the two operations is shown on the right (DIAS 1998; EUROPOEM 1997).

(Figure text:
Skin contact with pesticides during spraying with liquid products
milligram pesticide pr dag = milligrammes of pesticide per day
Fylde tank = Filling tank
Udsprøjte = Spraying
Fylde og udsprøjte = Filling and spraying
Indånding = Inhalation
Hænder = Hands
Krop = Body
Total = Total
Arbejdsoperationer = Work operations

The hands’ share of the total load is 62-94%. With powder products there is also relatively big exposure of the rest of the body. The exposure via the lungs is minimal compared with the total load. However, it should be noted that exposure via the lungs is normally more dangerous than exposure through the skin. The load averages 117 mg per day with liquid products, while the average exposure with powder products is 822 mg per day. Protection aids and clothing provide a low level of protection (50%), but if they are not used or are not functioning as they should, the exposure can be twice as big.

Figure 6.2
The sprayer operator’s exposure to powder products via inhalation, hands and body during different work operations (filling of tank and the actual spraying). The total exposure for the two operations is shown on the right (DIAS 1998; EUROPOEM 1997).

(Figure text:
Skin contact with pesticides during spraying with powder products
milligram pesticid pr dag = milligrammes of pesticide per day
Fylde tank = Filling tank
Udsprøjte = Spraying
Fylde og udsprøjte = Filling and spraying
Indånding = Inhalation
Hænder = Hands
Krop = Body
Total = Total
Arbejdsoperationer = Work operations

It is known that the potential exposure can be reduced by about 75% by using extra equipment in the form of, for example, preparation-filling equipment, hydraulic boom lift, hydraulic folding-in and out of the boom, non-drip valves, self-cleaning filter and tank-washing nozzle (Lund, Kirknel 1995). The yearly exposure of sprayer operators to pesticides depends on how many days per year they carry out spraying. Since the exposure per spraying day can be 500-4000 times greater than the average daily intake of pesticides via food (see section 6.2), the annual exposure of sprayer operators with many working days is considerably higher than the annual exposure via food products.

Exposure of greenhouse gardeners

For nurserymen working in greenhouses, exposure can occur during the spraying itself and in connection with work processes in the greenhouse after spraying. After spraying, the exposure depends on the time that elapses after spraying until the work in the greenhouse is carried out – the so-called re-entry time. During the work, the amount of pesticides and the area of leaves touched by the nurseryman during a working day are of critical importance. Important factors include the length of the working day and the substance’s degradation, evaporation and transferability from the treated plant surfaces to the nurseryman.

There are very few studies of greenhouse workers’ exposure to pesticides that can be used as models for exposure. Kirknel et al. (1997) mention a few German and some Dutch studies in this area. The result of the Danish studies accords with those of the German and Dutch studies. The Danish studies must be regarded as representative of a broad section of work functions in pot-plant nurseries in that they include not only work directly with plants but also moving of work-benches and packing of plants. Dutch studies in tall crops in greenhouses, such as tomatoes and cucumbers, fall within the result of the Danish model. The result of the Danish model is expressed as the 90-percentile, i.e. the highest exposure that 90% of a group of nurserymen can be exposed to. This cut-off point is generally regarded as high, with good safety. The Danish studies do not include the spraying of pesticides. Most spraying of pesticides in greenhouses is done automatically, with little risk of exposure. However, there is some limited spraying in Danish greenhouses with a hand-held spray gun, where the exposure is considered to be relatively high. Measurements of this work function are being studied in the UK (personal information, Erik Kirknel).

Kirknel et al. (1997) have developed a model for the exposure to and absorption of pesticides on re-entry into greenhouses. The daily potential exposure can here reach approx. 100 milligrammes of pesticide per working day, but typically lies in the interval 1-60 milligrammes in a working day of 6 hours with contact with the treated plants one day after the pesticide treatment. The most critical variable is the rate at which the pesticide disappears from the plant’s surface. In the exposure model, account is also taken of the amount of pesticide that can penetrate the skin. It is only the amount that penetrates the skin that causes possible health effects. The amount is normally less than the amount that lands on the skin.

6.1.4 Risk of cancer

Pesticides and cancer

Human data on exposure to pesticides and cancer are mainly based on occupational exposure in farming, forestry and horticulture and among workers employed in the production of pesticides. Reviews of these studies have not revealed a significant rise in the total mortality from cancer among persons occupationally exposed to pesticides (Dich et al. 1997; Maroni, Fait 1993; Blair, Zahm 1991). Furthermore, the total mortality was found to be lower among these groups of persons exposed to pesticides than among the general population. This has usually been ascribed to a "healthy worker effect" and, in the case of people employed in farming, to a healthier lifestyle among farming families. However, farmers, in particular, seem to have a higher occurrence of some types of cancer, including non-Hodgkin’s lymphoma, Hodgkin’s disease, multiple myeloma, leukaemia, soft-tissue sarcoma and cerebral cancer, skin cancer, lip cancer, stomach cancer and prostate cancer (Dich et al. 1997; Blair, Zahm 1995: Blair et al. 1992). Most epidemiological studies of cancer and pesticides have dealt with pesticides as a whole and are short on detailed information about the exposure. There are very few studies that have evaluated individual substances or classes of pesticides. Such studies are both difficult to design and difficult to interpret because people are seldom exposed to only one pesticide. Organochlorine compounds (to which the banned substance DDT belongs) have been associated with chronic lymphatic leukaemia, malignant lymphoma, multiple myeloma and soft-tissue sarcomas, and organophosphates and phenoxy acids have been associated with non-Hodgkin’s lymphoma. In cohort studies, insecticides as a group have been associated with increased risk of lung cancer, cerebral cancer and pancreatic cancer. A review of epidemiological studies of herbicides and cancer revealed reasonable evidence for assuming an association between non-Hodgkin’s lymphoma and phenoxy acid herbicides (Morrison et al. 1992). Furthermore, several studies found big increases in the risk of soft-tissue sarcomas from exposure to phenoxy acids but lacked proof of a dose-response relationship. Triazine herbicides have also been associated with non-Hodgkin’s lymphoma and soft-tissue sarcomas and with leukaemia, multiple myeloma, bowel cancer and ovarian cancer. However, in a review it has been found that the epidemiological data are insufficient to determine whether a relationship exists between exposure to triazines and cancer in humans (Sathiakumar, Delzell 1997).

Cancer in children of users of pesticides

There are only a few studies on the relationship between low-dose exposure to pesticides and cancer in children. The studies are based on different exposure scenarios – prenatal, postnatal, exposure in the home and parents’ potential exposure to pesticides in their occupation. Most of the data are from case control studies, and most of the research has been concentrated on leukaemia and brain tumours , presumably as an expression of the low occurrence of other types of cancer in children. The studies are limited by unspecific information on the pesticide exposure, potential recall bias, few cases, and most comparisons usually based on less than 10 exposed persons. However, many of the types of cancer that, in children, are associated with pesticides are the same types as are repeatedly associated with pesticide exposure among adults (Zahm et al. 1997), which could indicate a probable relationship. In addition, the observed risks are often greater among children than adults, which could indicate that children are more vulnerable to the carcinogenic effect of pesticides. Cerebral cancer is an example of a type of cancer in children that is frequently related to exposure to pesticides.

Other types of cancer that have been associated with exposure to pesticides include osteosarcoma, soft-tissue sarcoma, colorectal cancer, testicular cancer and other malignity in gametes, Hodgkin’s disease and retinoblastoma. With very few studies for each type, conclusions cannot be drawn about the possible significance of pesticides to the etiology of these types of cancer.

Limitations in the study of cancer in children

The studies generally suffer from methodological limitations. Incorrect classification of exposure, insufficient group sizes, bias in the choice of control persons and unchecked confusion are some of the main limitations in the case control studies of pesticides and cancer in children. Only a few studies have differentiated between the different groups of pesticides, and the exposure is usually dichotomised into having used pesticides at some time or another and never having used them, with regard for the frequency or duration of the exposure. The exposure data in all the studies are indirect and based on the parents’ job titles, occupation and use of pesticides in the home. In addition, it must be presumed that there is some degree of recall bias concerning details about the frequency and time of the use of pesticides in relation to conception, pregnancy and the child’s diagnosis – all things that can go back many years. The studies in which an association was found between pesticides and cancer seem to be those in which more detailed information had been received about the exposure with respect to timing, intensity or type of pesticide. Although several studies hint at an association between pesticide exposure and certain types of cancer, there does not seem to be sufficient epidemiological evidence of an etiological relationship between exposure to pesticides and cancer in children (Daniels et al. 1997; Zahm, Ward 1998).

Studies of the risk of cancer in connection with the use of pesticides are often complicated by the fact that the individual farmer uses many different kinds of spray products and that the individual pesticides have different toxicological profiles. Many studies have raised a suspicion of pesticide exposure as the possible cause of increased frequency of certain form of cancer in the blood and lymphatic tissue (Zahm et al.1997). Genotoxic damage, as seen in non-Hodgkin’s lymphoma patients, has also been found in lymphocytes from peripheral blood in persons exposed to pesticides (Garry et al. 1996). In this connection, it must be mentioned that farm workers have been exposed many times. Such repeated exposure may also affect the individual farmer’s immune system (Blair, Zahm 1995).

In Denmark there have not been any real epidemiological studies concerning pesticide exposure of farm workers, but several researchers have studied pesticide exposure in horticultural workers. There have been a few studies of "mortality and occupation".

Studies of frequency of cancer

An analysis of cancer cases related to occupation in the years 1970-79 showed an over-frequency in agriculture of some forms of cancer in the blood and lymphatic tissue. In the case of men employed in agriculture, 22 cases of acute leukaemia were found, against an expected 12 cases. In addition, a significantly increased risk of chronic leukaemia was found among men in agriculture, with 32 cases of non-acute leukaemia against an expected 19.2 cases. In this study, there were considerably fewer cases of lung cancer among agricultural workers than expected (Olsen, Jensen 1987). There is no follow-up on this study from the period after 1979.

6.1.5 Other effects

Studies in this section have been discussed by Skadhauge (1998). However, the best-documented studies of pesticides were carried out with substances that are no longer used or that have never been used in Denmark.

Reproductive toxicity

It has been documented that occupational exposure to pesticides can have a negative effect on fertility (Smith et al. 1997; Strohmer et al. 1993; de Cock et al. 1994). A known example is the substance dibromide-chloropropane (DBCP), which caused azoospermia and oligospermia among Californian workers working with the substance (Whorton, Foliart 1983). Other pesticides, such as ethylene dibromide, kepon and carbaryl have been associated with reproductive effects in males (Baker, Wilkinson 1990). In some studies an association has been found between miscarriage and foetal death and occupational exposure to pesticides (Pastore et al. 1997; Goulet, Thériault 1991), whereas other studies have been unable to demonstrate such a relationship (Restrepo et al. 1990; Willis et al. 1993; Kristensen et al. 1997a).

In a review article from 1995 it is concluded that there is no clear epidemiological evidence of a relationship between exposure to pesticides and increased reproductive risk (Nurminen 1995). A large Norwegian study of congenital deformities in children born of parents that were registered as farmers found an association between pesticides and deformed sex organs (Kristensen et al. 1997b). A large review article was published recently on studies concerning potential associations between foetal deaths, miscarriages and stillbirths and specific pesticides, together with parents’ employment in occupations with potential exposure (Arbuckle, Sever 1998). Data indicated an increased risk of foetal death associated with pesticides in general and the mother’s employment in agriculture. However, it was concluded in the review that the studies carried out to date do not answer the question concerning the toxic effect of individual pesticides on human reproduction.

Effects on reproduction

A study of greenhouse workers showed significantly reduced plasma-cholinesterase activity compared with an unexposed control group (Lander et al. 1995). In a study of semen quality and chromosomal damage in greenhouse workers exposed to pesticides, no link was found between individual factors, including exposure to pesticides. That applied both to specific linkage with measured exposures to pesticides and broad linkage with the market garden’s use of pesticides (Abell et al. 1997). The most important observation was that both chromosomal damage and sperm quality were related to the current pesticide exposure and that spraying was less important than exposure on re-entry. The study revealed no differences between the greenhouse workers’ sperm quality and the sperm quality of organic cultivators. In addition, the greenhouse workers had a generally higher sperm quality than the general population. On the other hand, it was observed that the longer the persons studied had worked in horticulture, the poorer the sperm quality. However, this was not unambiguously correlated with pesticide exposure, nor was a correlation found between the pesticide consumption of the market garden in question and sperm quality, although the workers with a low exposure had a better sperm quality than those with a high exposure. The results of the study indicate a need for increased action to reduce exposure of greenhouse workers when handling sprayed plants.

Developmental toxicity

Several studies have shown developmental effects as a consequence of parents’ occupational exposure to pesticides. In the above-mentioned Norwegian study of congenital deformities among newborn children of parents registered as farmers, a moderately increased risk of spina bifida and hydrocephalus was found, compared with children born of parents in other occupations in rural communities. The risk was greatest in the case of exposure to pesticides in orchards and greenhouses (Kristensen et al. 1997b). Exposure to pesticides, particularly in the case of arable farmers, was also associated with limb defects. A Dutch study from 1996 showed an increased risk of spina bifida in children born of mothers employed in farming, compared with a control group, but the association could not be explained by use of pesticides (Blatter et al. 1996). However, in a Finnish study of congenital deformities and mothers working in farming, the risk to workers exposed to pesticides was found to be no greater than the risk to unexposed farm workers (Nurminen et al. 1995). A recent review article describes methods and results of studies of occupational exposure to pesticides, mainly among farm workers, and the risk of congenital deformities. However, on the basis of the available information, there seems to be insufficient evidence to date to either confirm or disprove a relationship between exposure to pesticides and deformities (García 1998).

Neurotoxicity

With respect to neurotoxic effects of pesticides in adult populations, it has been found in several studies of workers exposed to pesticides that effects can occur in the peripheral nervous system in workers with either acute poisoning or with chronic occupational exposure without obvious neuropathic syndromes (Keifer, Mahurin 1997; Ecobichon 1996). Most of the studies of the cognitive effects of exposure to pesticides have concerned organophosphates because of their widespread use.

Following acute exposure to high doses of organophosphates, with repeated, acute, clinically significant intoxication, toxic effects have occasionally been observed with long-term effects on behaviour and on mental and visual function (Rosenstock et al. 1991; Ames et al. 1995¸Steenland et al. 1994). However, the available data do not indicate that asymptotic exposure to organophosphates is associated with an increased risk of delayed or permanent neuropsychopathological effects (Daniell et al. 1992; Eyer 1995).

Among fungicides, the dithiocarbamats have been associated with neurotoxicity in a few cases. In a study from the Netherlands, both autonomous and peripheral neurotoxic effects were found among workers chronically exposed to zineb and maneb in flower production (Ruijten et al. 1994). These pesticides are not themselves suspected of being peripheral neurotoxins, but carbon-disulphide, which is one of the metabolic products, is a known neurotoxin. Furthermore, occupational exposure to pesticides containing manganese has been mentioned as a possible cause of manganese poisoning of the central nervous system (Ferraz et al. 1988).

In an epidemiological study from Calgary, Canada, persons with earlier occupation exposure to herbicides were found to have a three time greater risk of Parkinson’s disease (Semchuk et al. 1992).

Immunotoxicity and sensitivity

In an American study of 280 cases of aplastic anaemia, an association was found with occupational exposure to organochlorine compounds and organophosphates (Fleming, Timmeny 1993). It can be concluded that there is evidence of contact hypersensitivity as a consequence of occupational exposure to pesticides.

Experimental and clinical data have shown that some pesticides (chlornitrobenzene, carbamats, captan and organophosphates) can induce contact hypersensitivity (type IV reaction) in test animals and humans (Baker, Wilkinson 1990).

Mortality among farmers

In the third report from the Ministry of Health’s Life-expectancy Committee, standardised mortality ratios (SMR) were given for different occupations. For each occupation, the causes of death were also examined. The period covered was 1986-90. Compared with largely all other occupations, "independent in farming" had the lowest SMR (Ingerslev et al. 1994).

6.1.6 Conclusions

The risk of acute effects from pesticides is deemed to be considerably lower today than it was just 10 years ago. With use of the protection aids that are recommended for the individual pesticide according to its classification and labelling, there is a minimal risk of incurring chronic health problems. The possibility cannot be excluded of some risk to persons who do not observe the given rules for personal protection and correct use of the pesticides, inappropriate work routines and poor work hygiene. However, the sub-committee notes that there can be considerable exposure of the sprayer operator and of workers in greenhouses and in the production of fruit and vegetables, where frequent use is made of pesticides.

In tractor work in the field, whole-body vibrations occur. The tractor driver also has to twist his back many times. He often needs to look behind, whereby his spinal column, neck and shoulders are loaded. Farmers have a generally increased risk of osteoarthritis, which is associated with milking, tractor work and heavy physical work, which are often started before the age of 16 years. Persons working in dusty conditions have an increased risk of asthma and chronic bronchitis.

Farm workers are exposed to noise, partly through work in stables and partly from tractors and other agricultural machines.

Pesticides are used in farming and horticulture to control weeds, fungi, insects and other pests and to influence the growth of crops such as fruit, vegetables and cereals. In Denmark, there are at present about 90 authorised active ingredients, which are used in approx. 550 products.

Over the years, the requirements for authorisation of the individual pesticides have been considerably tightened. As a result, some pesticides have been banned on account of their undesirable environmental and/or health properties. This applies, for example, to a number of pesticides containing chlorine, such as DDT, dieldrin, etc. However, owing to their lack of degradability, they are still found in the environment, from which they find their way into food products.

Pesticides are widely used. Within the food sector, it is particularly in connection with the production of fruit and vegetables that the risk of a residual content is greatest.

Extensive use is also made of pesticides, in the form of herbicides and fungicides, in the cultivation of cereals and to regulate growth. The crop is often sprayed long before it is harvested, so the degradation of the pesticide can be advanced. When a crop is sprayed a shorter time before harvest, it must normally be expected to have a residual content of the pesticide.

Acute and chronic effects of pesticides have been investigated in numerous experimental studies and in studies of persons with occupational exposure, both in Denmark and in other countries. Apart from high-dose exposure in connection with accidents or suicide, the general public in Denmark is subject mainly to low-dose exposure, primarily as a consequence of pesticide residues in food and drinking water or from private use of pesticides in or around the home.

Population studies

There are only extremely sparse epidemiological data concerning health effects among the general population as a consequence of low-dose exposure to pesticides. There are differences in both exposure and sensitivity to pesticides in the population, depending on age, sex, eating habits, environmental factors and/or lifestyle. Some sections of the population must thus be expected to show health effects from exposure to significantly lower doses of pesticides than those causing effects in the rest of the population. Children are a special risk group, particularly due to qualitative and quantitative differences between children and adults in their intake of different types of food. Other risk groups may be persons with a poor immune system or persons with certain chronic diseases. For further amplification of the effect of pesticides on public health, readers are referred to Skadhauge (1998).

Metabolites

In living organisms, pesticides are transformed into metabolites. The documentation on which the authorisation of pesticides is based normally includes information on most of the metabolites from the substance in question, but separate toxicological investigations of the individual metabolites are not normally performed. The toxic effect of metabolites is generally thought to be lower than that of the original substance because the metabolites are often more water-soluble and are thus eliminated faster from the organism (Hayes, Laws 1991). However, there are some important exceptions, where biotransformation results in a more toxic product (e.g. the formation of oxon derivatives of organothiophosphorous pesticides or the formation of epoxy compounds from certain insecticides, such as dieldrin and heptachlor). Atrazine’s metabolites, deethylatrazine and desisopropylatrazine, are more acutely toxic than atrazine; they leach more easily than the original pesticide and may thus be a bigger problem. Where there is knowledge about such metabolites, it is taken into account in the authorisation procedure and in connection with the setting of limit values in food.

6.2.2 Risk population and individual vulnerability

In the case of exposure to a potential toxic factor, the risk population is traditionally taken to be the part of the population that is subject to 1) increased exposure, 2) increased dose with the same exposure, 3) increased effect with the same exposure in relation to the rest of the population.

Increased exposure can occur in the case of accidents or occupational exposure, but it can, for example, also occur to some extent in population groups with special eating habits (e.g. vegetarians or persons with another food base) or neighbouring on sprayed land.

Children as a risk group

Children are a risk group because they are often exposed to a larger dose of pesticides than adults with the same type of exposure. Owing to more rapid breathing in relation to their body weight than adults, children inhale relatively more air. Children’s special diet patterns and intake of pesticide residues are the subject of a report from the American National Research Council (NRC). Here, both qualitative and quantitative differences were found between children and adults with respect to exposure (National Research Council 1993). Firstly, children take in more energy per kg body weight than adults and, secondly, children eat far fewer types of food than adults. In addition, the intake of water, both in the form of drinking water and as a food component, differs greatly between children and adults. The council concluded that differences in diet and thus in dietary exposure to pesticide residues could explain most of the differences in pesticide-related health risks between children and adults. Differences in exposure were generally a more important cause of differences in risk than age-related differences in the way the substances act in the organism. However, the council found that, among other things, effects on neurological and immunological development processes were insufficiently clarified.

With respect to an increased effect with the same exposure to pesticides, there is agreement that foetuses and children are a special risk group (Goldman 1995; Reigart 1995). As stated in the above-mentioned report from NRC, there are both qualitative and quantitative differences between children and adults with respect to the toxicity of chemical substances, including pesticides (National Research Council 1993). The report gives examples, mostly concerning medical drugs, where children are more or less sensitive to the individual substances. These differences result from the fact that the substances are degraded at a different rate in the body in children than in adults and from the greater or lesser toxicity of the degradation products compared with the parent substance. The qualitative differences in toxicity are due to exposure in particularly sensitive periods in early development, when exposure to a toxic substance can permanently change the structure or function of an organ system. The quantitative differences in toxicity between children and adults are due partly to age-related differences in absorption, metabolism, detoxification and excretion of the environmentally foreign substances and partly to differences in size, not fully developed biochemical and physiological functions and variations in the composition of the body (water, fat, protein and mineral content), all of which can affect the degree of toxicity. Since new-born infants are the group that differ most from adults, anatomically and physiologically, they must be regarded as having the most pronounced quantitative differences in sensitivity to pesticides. The report found that quantitative differences in toxicity between children and adults were normally lower than a factor of 10.

With reference to the foregoing, the US Environmental Protection Agency (US-EPA) estimates that children constitute a special risk group. Accordingly, in individual cases in the USA, an extra uncertainty factor of 100 is used when setting maximum limit values in connection with risk analyses that are based on animal tests and that are not deemed to throw sufficient light on the special situation of children. Examples of such substances and the related uncertainty factors (in brackets with the year) are: heptachlor (200 in 1990), triazophos (500 in 1990), fentin (200 in 1970; 500 in 1991), abamectin (500 in 1992), amitrol (provisional uncertainty factor of 1000 in 1993), phosalon (200 in 1993) and propylene thiourea (metabolite of probineb, 1000 in 1993).

Other risk groups

Pregnant women, the elderly and persons with a poorly functioning immune system are other groups that must be presumed to be particularly at risk with respect to health effects from exposure to pesticides.

6.2.3 Exposure of the population

The general population can be exposed to pesticides through their intake of pesticide residues in food and via drinking water or through use of pesticides in and around the home, schools, offices, public parks and farms (drift and evaporation). Exposure can also occur through accidents, leakages due to incorrect storage, spills and pollution from production plants or landfill sites. Children, in particular, can be exposed to pesticides from eating contaminated soil, from contact with pets treated with pesticides or from parents exposed to pesticides in their work.

Exposure in the home

Exposure to pesticides can occur in connection with their use in and around the home, e.g. in combating insects and treating pets. Besides direct indoor treatment with pesticides, pesticides can be given off by pot plants treated with pesticides at the nursery. There have been no studies of this exposure in Denmark (Skadhauge 1998).

Exposure via food

The main exposure of the population to pesticides must be assumed to occur via food. Around 60% of this intake occurs via imported food products. Exposure via vegetables and, particularly, via berries and fruit predominates. The sources of the information on this are the Veterinary and Food Administrations’ annual reports on the nationwide control and monitoring of the residual content of pesticides in vegetable and animal food products on the Danish market. The samples are extracted at the wholesale level. The studies cover both Danish and imported food products and are carried out as a random sample control, supplemented by targeted control.

Figure 6.3
Frequency of fruit and vegetables with pesticide residues for Danish and imported crops, 1993-1997 (fruit and vegetables with the highest number of samples are presented). On average, pesticide residues have been detected in about one quarter of the Danish and imported fruit shown in the figure. For the vegetables, pesticide residues have been detected in an average of 20% of the imported vegetables in the figure, compared with 6% of the Danish vegetables.

Figure text:
frugt = fruit
æbler = apples
solbær = blackcurrants
ribs = redcurrants
pære = pears
kirsebær = cherries
jordbær = strawberries
hindbær = raspberries
blomme = plums

grøntsager = vegetables
tomat = tomatoes
salat = lettuce
kinakål = Chinese cabbage
kartofler = potatoes
gulerod = carrots
courget = courgette
bladselleri = celery
agurk = cucumber
importeret = imported
dansk = Danish

Food products

In the national monitoring programme, the samples are extracted at random (randomised) from the crops that it has been decided to check in the year in question. The programme is designed to determine the level of residues in the crops checked. The samples are taken at wholesalers, producers and importers. The sampling is carried out by the food inspection units, together with the Plant Directorate and the Danish Veterinary and Food Administration. The samples are generally extracted with a geographical distribution and distributed over the whole of the year In the case of fruit and vegetables, samples are taken weekly. In 1997 1,947 samples were extracted – 83% fruit and vegetables (1,613 samples), 3% honey (Danish and imported), 8% meat, 6% cereal (including 1/3 imported cereal). The 1,613 samples of fruit and vegetables comprised 919 samples of imported products and 694 Danish products. In both 1996 and 1997, the monitoring programme constituted about 0.0002% of the Danish production of vegetables, fruit and berries.

The national monitoring programme

In the last couple of years, studies of fruit and vegetables have covered about 150 different pesticides, comprising commonly used insecticides and fungicides and some herbicides. When the substance profile for the programme is set, the substances that have Danish MRL-values (MRL = Maximum Residue Limit; often corresponding to EU MRL-values) and/or are used here in Denmark are prioritised. Otherwise, the pesticides with low MRL-values or that are widely used are prioritised. Animal products and fish are checked for organochlorine pesticides, which used to be authorised for use and now occur as widespread environmental pollution. At the same time, the samples are tested for the industrial chemical PCB. In addition, a number of pesticides containing chlorine occur in fish – particularly oily fish. Since these substances have long since been banned, they are regulated and treated as environmental pollution. The occurrence of these substances in fish would not be affected by a reduction or possible phasing out of pesticides in Denmark.

Special investigations (targeted control)

In addition, special investigations are carried out, including so-called targeted control, that is used where and when it is suspected that current MRL-values are being exceeded. The suspicion can, for example, arise in connection with the current monitoring programme. The nature and scope of the targeted control vary from case to case and comprise two or more samples extracted at the retail and/or wholesale level.

Exceedances of limit values

For many years, the analyses carried out have shown that pesticide residues in food products on the Danish market generally meet the current regulations. The content has been found to exceed the MRL-values in 1-2% of the samples, but rarely by more than 50-100%. It is characteristic of the analyses that, with the reporting limits used, no residues were found in most of the samples. Normally, residues are detected in one third or less of the samples tested. As shown in figure 6.3, the detection frequency is higher in samples of imported food products than in Danish food products, but within the individual crops, the detections vary, both qualitatively and quantitatively from year to year, depending on economic and climatic conditions and on differences in the need for treatment (Büchert 1998). In 1996, the most frequently detected pesticides were (in alphabetical order): captan, carbendazim, chlorothalonil, dithiocarbamates, endosulfan (sum), iprodione, quintozene, tolylfluanide and vinclozolin. None of the finds gave the Veterinary & Food Administration cause for concern with respect to public health (Büchert, Engell 1998).

Finds and exceedances in relation to crops

The monitoring programme indicates a tendency towards relatively more detections of residues in fruit, including citrus fruit and grapes, than in vegetables such as cabbage and potatoes. The total human average intake of pesticides is dominated by citrus fruit in particular (oranges and mandarins), potatoes and apples and, to a slightly lesser extent, tomatoes, pears, grapes and strawberries, which account for most of the load. However, the sum of these substances covers pesticide residues with very different toxicological properties, so the substances are not directly comparable and their combined action is not known. The Ministry of Food, Agriculture and Fisheries has initiated a study that will throw light on these questions within some years. Another important factor in connection with the summation of the pesticide residues in food products is that it does not include reduction factors. Some of the pesticide residues are removed during preparation - for example, peeling of oranges, mandarins, etc., the peel of which is seldom used. However, there are as yet insufficient data to clarify the size of the reduction factors in Danish diet. This aspect will also be looked at in a new study initiated by the Ministry of Food, Agriculture and Fisheries. A closer comparison of the individual food products and their significance for the intake of pesticides, and thus their possible toxic effects, is difficult for several reasons, including the fact that the data are often too limited, but also because the individual pesticides have different toxicological properties. It must be mentioned, however, that the higher detection rate in imported crops reflects the fact that, for comparable crops, there tend to be fewer positive results in Danish products than in imported ones. This tendency is most pronounced for crops like tomatoes, cucumber and peppers, which are greenhouse crops in Denmark but outdoor crops in other countries. However, the tendency does not apply to all crops. There are also examples, e.g. blackcurrants, where pesticide residues are more frequent in Danish crops than in imported ones.

Calculation of pesticide intake from food products and drinking water on the Danish market

The following calculation of the Danes’ intake of pesticides through their diet is based on monitoring data from the Danish Veterinary & Food Administration’s national monitoring programme in 1996 and 1997 (Büchert 1998). These data are not only the latest data but were also obtained with comparable methods and reporting limits. It is characteristic that no content was detected in most of the samples tested. Regardless of this, it must be assumed that treated crops always have a certain residual content, so that lack of detection can only be taken to mean that the content, if any, is below the analytical detection limit.

Methods of calculation

Different models can be used to calculate the pesticide intake from crops in which no residual content has been detected. In situations in which there are sufficient data above the detection limit, the distribution under the detection limit can be determined with reasonable statistical certainty. Unfortunately, that is not the case with the available monitoring data. The residual content in samples without a detected content must therefore be expressed in a different way – for example, by putting the pesticide content at 0 (zero) or, conversely, by assuming that it corresponds to the detection limit. However, both are rather rough approximations.

In this connection it must be stressed that the analytical detection limit is not a fixed quantity. It varies from one pesticide to another and from crop to crop, just as there is a variation from one analysis to another of the same substance in the same crop. However, since the exact data are not directly accessible, the calculations are based on the more general detection limits, which are determined for the individual substances during the validation of the methods. The residual concentration in samples with concentrations above the detection limit are estimated as the mean value between the highest and lowest detected content.

The scope of the calculations

The calculations of the content in the crops analysed have been carried out for all the pesticides detected in the analyses in 1996 and 1997. It was anticipated that there was a content in both Danish and foreign crops of the same type, also when the substance was only detected in crops of one or the other origin. There is also an overestimation, in that the calculations are based on an assumption that detection of a pesticide in a single type of crop means that the entire production of the crop in question has been treated with the pesticide in question and therefore has a residual content of the pesticide that differs from 0.

The calculations do not take into account intake of pesticides that have not been detected in the national monitoring programme in either Danish or imported crops, nor do they include any intake of such substances as might occur in contents below the analytical detection limits or that are not covered by the analytical methods used.

In the calculation of Danish consumers’ exposure to pesticide residues through food, the crops’ estimated content is multiplied by the average daily dietary consumption of the crop in question. The dietary data used are based on diet studies from 1985 and 1995 and on trade statistics.

Variations in the population

In practice, of course, there are many deviations in the population from the average diet model with respect to sex, age, ethnic background and social conditions. For a more detailed risk analysis for such groups there is thus a need for more precise dietary data. However, it must be stressed that the model set up provides a reasonable possibility of a more general evaluation of the pesticide intake on a national level, including the broader significance of keeping Danish crops free of pesticide treatment.

The distribution between Danish and imported food products

Good data on the distribution of Danish consumption between Danish and imported crops are not available. The main reason for that is, that with the "internal market", official statistics are no longer kept of imports and exports between the Member States. In the calculations carried out, the distribution has been estimated on the basis of earlier trade figures and agricultural statistics. These indicate that the distribution between imports and domestic production is 1:1 for such fruit as apples, pears, plums, berries, etc., while exotic fruit, such as citrus fruit and kiwi are exclusively imported crops. For vegetables, the distribution is about 1:4, although not for cucumber, tomatoes and similar, where imports, distributed over the whole year, account for about 70% of the total consumption. In the case of cereals, maize and rice are exclusively imported crops, while the consumption of barley is based entirely on Danish products. In the case of rye, wheat and oats, 5%, 20% and 65%, respectively, of consumption is covered by imports.

Table 6.2 shows the average daily consumption of Danish and imported cereal and cereal products in Denmark.

Table 6.2
Average consumption of Danish and imported cereal and cereal products, given in grammes per day.

The Danes’ dietary pattern

The variation in the Danish consumers' dietary pattern can be judged on the basis of the National Food Agency's dietary study in 1995. The results of that study, which covered more than 1,800 persons, are summarised in table 6.3 below, which gives the average intake and selected fractiles for the adult part of the population.

Table 6.3
Variation in Danish diet, given in g/day for adults m/f (Büchert 1998).

On the basis of the values in table 6.3 it can be calculated that the 90% fractile for dietary consumption of vegetables is about 225% of the mean intake, while, for fruit, it is about 300% and for cereal products, about 200%. The uncertainty on the figures is reflected by the fact that the values for the 90% fractiles are of the same order of magnitude as the sum of the mean content and the standard deviation. It must be added that the probabilities have not been calculated for how population groups and individuals compose their dietary consumption.

6.2.4 Calculations of the population’s intake of pesticides

Reduction factors and adjustments

The intake calculations for fruit and vegetables, which, as described, are based on monitoring data, reflect the occurrence of pesticides in the raw produce, so account is not taken of the reduction of the content during preparation of the raw produce (peeling, boiling, etc.). This is because the literature contains only limited information about such reduction factors and the data are usually of such a nature that they cannot be applied to Danish conditions. Therefore, reduction factors have not been included in the current calculations of the intake of fruit and vegetables, which leads to overestimation of the intakes in relation to the real values. A lack of data on the proportion of a given crop that has been treated with a specific pesticide has also made it impossible to adjust for actual use. This applies particularly to imported products and means that, for example, the intake from citrus fruit – and thus from imported crops – has been overestimated.

Estimated intake from fruit and vegetables

The intake of pesticide residues through diet has been calculated for all the pesticides detected in the analyses of fruit and vegetables in 1996 and 1997. The intake from Danish crops and from imported crops has been calculated separately, without detailed specification of the origin of the crops (for details, see Büchert 1998). The results show a total average intake of pesticides from fruit and vegetables of about 165 microgrammes per day. The intake of six pesticides/pesticide groups – carbendazim, dithiocarbamats, iprodione, o-phenyl-phenol, procymidone and thiabendazole – corresponds to half the total intake, while the other half is distributed over about 60 individual compounds.

The calculated total intake of pesticides from fruit and vegetables is made up of around 165 microgrammes per day is approx. 60 microgrammes per day, corresponding to 36%, from Danish crops and approx. 105 microgrammes/day, corresponding to 64%, from imported crops.

It must be stressed that any size comparison of the calculated intakes should be seen in the light of the fact that the pesticides have different toxic properties and potency. A small intake of a potent substance may thus very well be more dangerous from a health point of view than a large intake of a less potent substance.

Comparison with ADI

For the pesticides that are authorised for use on edible crops, an acceptable daily intake (ADI) is fixed on the basis of the same principles as for additives, and with a requirement concerning similarly comprehensive data. ADI for pesticides is set to protect against possible long-term effects. ADI is fixed on the basis of animal tests. In each test, NOAEL (No Observed Adverse Effect Level) is fixed as the dosage that does not have any harmful effect. ADI is fixed by taking the NOAEL from the most sensitive test on animals and reducing this by an uncertainty factor that must take account of the uncertainty that lies in extrapolating from animals to humans and the variations in people’s sensitivity and lifestyle. By international agreement, an uncertainty factor of 100 is used as the basis.

A comparison has been carried out between the calculated intakes and the upper limits recommended by the experts for acceptable intakes for a person weighing 70 kg, calculated as the ADI values multiplied by the weight (ADI mg/kg bw/day times 70 kg). All the calculated intakes constitute only a small part of the intake that could be accepted without giving rise to health concerns. The mean value of the calculated intakes of the individual pesticides is 0.31% of the upper limit for their acceptable daily intakes, with a standard variation of 0.46% and with the highest individual value (for methidathion) of about 2.2%. Even with big variations in the dietary consumption of the individual crops, the consumers would remain below the ADI values.

The variations in the exposure to pesticides via fruit and vegetables have been clarified by calculating the intake at the 90% fractile. The 90% fractile for vegetables and fruit has been put at 225% and 300%, respectively, on the mean value for the group of crops in question.

An evaluation of the maximum exposures of the population can best be based on the calculated mean intakes and the values given in table 6.3 for the entire group of vegetables or fruit. An evaluation of the intake via the individual crops can also be based on the data, whereas a summation of the maximum contribution from the individual crops is not possible without calculating the probability of the same person or population group having maximum consumption of several combinations of types of crop and determining the types of crop in question. On the face of it, it is believed that a conjuncture of maximum consumption of two or more crops would be exceptional and that a major consumer of a crop would normally have a limited consumption of other types of crop. In other words, if there were a maximum exposure to pesticides via one type of crop, the exposure to the others would often be correspondingly smaller.

Estimated intake from cereals and cereal products

Since 1987, cereals and bran have been included in the regular analyses of pesticide residues in food products on the Danish market. The analyses have been carried out with a changing search profile from year to year. In the last couple of years, the samples have been analysed with a multi-method covering 24 different substances. In 1997, a special analysis was carried out of the residual content of the straw-shortening products chlormequate and mepiquate. The analyses showed that the residual content of pesticides in cereals and cereal products meets current limit values, although, generally speaking, the data are too slender for a more refined analysis of the population’s exposure to pesticides from products of this type. The highest content of chlormequate – 3.8 mg per kg – was found in a sample of rolled oats from the UK. There are generally more and higher finds in Danish cereal products than in the imported products, and pesticides have been detected in almost all samples except in rice and in a sample of organic wheat. For a detailed review, readers are referred to Granby and Poulsen (1998). The calculations of the intake were based on a more conservative estimate of the residual content in the types of crop in question. The calculations were carried out in the same way as for fruit and vegetables, but covered all the pesticides detected in the entire period from 1987 to 1997.

Reduction factor

In the calculations, no direct differentiation was made between whether the residual content was detected in bran, in grain or in flour. If a substance was detected in one type of product, it was assumed that the substance was also present in others. It will normally be bran that has the largest residual content, and the residual content in flour will normally be reduced when the husks are removed and be reduced still further when the grain is milled and when the flour is baked into bread (bread was not included in the analysis). On the other hand, most of the pesticide residues will normally be in the bran product. In the case of chlormequate, it was found that milling of wheat grain into flour reduced the residues by a factor of about 4. For this reason, in the calculation of the intake from cereals and cereal products, it was assumed, as a cautious estimate, that the residual contents were reduced by a factor of 2 in relation to the measured content. In 1997, 21 bran samples were extracted (18 wheat bran and 3 oat bran), which were analysed for 26 pesticides. None of the samples exceeded the limit values, even though the residual contents were higher in bran than in whole grains and flour. However, reservation must be made for the fact that the analyses for glyphosate were of limited scope.

The results of the calculations are summarised in table 6.4, which shows the calculated pesticide intake with and without reduction through processing of the cereal. As will be seen from the table, the average intake is estimated to be about 26 microgrammes per day, with a distribution of about 2/3 from cereals produced in Denmark and 1/3 from imported cereals.

Bromide from fumigation with methyl bromide

It should be noted that the residual content of bromide is not included in the summation of the total intake. Bromide is not in itself a pesticide but is included in the analyses as an expression of any use of methyl bromide, which is used to fumigate cereals and other products. The test method for measuring methyl bromide is based on determination of bromide, but as this also exists in nature, it is a question of measuring whether the bromide content is higher than the normal content in the cereal product. In other words, the absolute content of bromide does not correspond to the residual content of methyl bromide, and it would be a gross over-estimation to include the bromide content when summing the total intake.

As in the case of fruit and vegetables, the average intake of the individual pesticide residues from cereals and cereal products is typically less than 1% of the ADI values (see Büchert 1998).

The variations in the exposure to pesticides from cereals and cereal products can be estimated from table 6.3, from which it will be seen that the 90% fractile for cereals and cereal products corresponds to approx. 200% of the mean intake.

Table 6.4 Look here!
Estimated pesticide intake from imported and Danish cereals and cereal products (Büchert 1998). There are very few analyses of glyphosate in cereals and cereal products. Empty spaces indicate that residues have not been detected. The analyses included pesticides that were not detected in any of the samples.

The estimated intake from animal food products and fish

The Danish Veterinary & Food Administration has been monitoring the content of pesticide residues in animal food products for many years. The analyses have primarily been directed at the chlorinated pesticides, such as DDT, HCB and similar fat-soluble compounds, but analyses have also been carried out for residues of organophosphorous pesticides in meat.

No residual content of organophosphorous pesticides has been detected in meat, but a chronic content of chlorinated pesticide residues has been found in both animal food products and fish. However, this residual content represents substances that have been banned in pesticides for many years. Their occurrence in the food products is therefore a result of their earlier use, which has resulted in extensive and permanent environmental pollution with these persistent substances.

Since the exposure of the consumers to pesticide residues from animal food products and fish is primarily a question of "the sins of the past" and does not concern the pesticides authorised for use today, the exposure does not depend on the extent of the use of the pesticides currently used in Denmark.

The estimated intake from drinking water

In the last few years it has been recognised that there is widespread pollution of our groundwater with pesticides. There is therefore a potential risk of such contaminants in our drinking water and thus a risk of exposure of the population through the water.

Since drinking-water wells in which a residual content has been detected have hitherto been closed, the real exposure of the consumer is limited. However, this does not exclude the possibility of undiscovered pollution of local wells that can result in significant exposure of the local inhabitants.

It must therefore generally be assumed that drinking water meets current limit values, which state that the content of individual pesticide residues must not exceed 0.1 microgramme per litre, while the total content of pesticides must not exceed 0.5 microgrammes per litre.

With a normal consumption of 2 litres of water per day, a consumer thus receives less than 0.2 microgrammes of an individual pesticide, and his or her total pesticide intake from drinking water is less than 1 microgramme per day.

When the estimated maximum intake from drinking water of 1 microgramme per day is compared with the estimated intake of approx. 190 microgramme per day from fruit and vegetables, it must be concluded that the intake from drinking water is generally negligible and disappears in the uncertainty that lies in the estimation of the intake from vegetable food products.

Overall assessment of intake from food products and drinking water

As will be seen from the foregoing review of the pesticide intake from the different types of food products and drinking water, the predominant sources of the general population’s exposure to pesticides are fruit and vegetables and cereals and cereal products, while the intake from drinking water, animal food products and fish is of no significance for the total exposure.

The total average exposure from the individual types of food product and drinking water is summed in table 6.5.

Table 6.5
Estimated average pesticide intake from Danish and imported food products (Büchert 1998).

As will be seen from the table, the total average intake is estimated to be approx. 200 microgrammes of pesticide per day, about 60% of which comes from imported products, and the remaining 40% from Danish products, including cereals.

Variation in daily intake of pesticides

A rough calculation can be made of the variation in the intake of pesticides. These calculated extremes must be treated with caution because a closer statistical calculation is expected in the future. The lowest intake of pesticides is among people who either have a very low consumption of vegetable food products or who base their consumption solely on untreated, possibly organic, products. On the high side, the extreme must be presumed to be a so-called vegan, who bases his or her entire energy consumption on traditionally cultivated fruit and vegetables. In all probability, there are not many 100% vegans – people who live exclusively on vegetable products. A vegetarian typically covers part of his or her energy need with dairy products and eggs. For this reason, there are no detailed data on the typical composition of such vegans’ diet, but a total intake of fruit and vegetables and cereals and cereal products of around 2½ kg per day must clearly be regarded as at the upper end of the real consumption. That corresponds to three times the average consumption on which the calculations of the average Dane’s intake of pesticides through their diet are based. Judged in this way, the extremes for pesticide intake are estimated to lie from a very low intake to about three times the calculated average intake, corresponding to 570 microgrammes per day. Although this is an estimated calculation of the upper limit, it accords well with the calculated intake for the 90% fractile of the population.

The intake in relation to ADI

The calculations also show that the average intake at individual substance level is typically about 1% or less of the acceptable daily intake (the ADI value) for the individual compounds. Even with big variations in food consumption, the intake thus lies below the acceptable levels. As shown in figure 6.4, there are seven pesticides with a residual content of more than 1% of the ADI value for the pesticide in question, while all the other pesticides lie below 1% of ADI (Büchert 1998). This figure does not necessarily show the seven most eaten types of pesticide residues, which is illustrated by the fact that according to the diet data used, more than 90% of the two pesticides ortho-phenylphenol and methidathion are received through citrus fruit. Citrus fruits are usually peeled before being eaten, and there are today experiments that show that methidathion in citrus fruits is reduced by 95% by peeling the fruit. That means that the figures for methidathion, and presumably also for ortho-phenylphenol, can be more than 80% too high. Other pesticide residues are found mainly in crops that are not peeled or processed (e.g. cucumber, apples and pears), and the intake of these pesticides could in reality be more important than the real intake of the pesticides shown in figure 6.4. It must be stressed that it cannot be absolutely claimed that all calculations that do not include reduction factors are too high, since during preparation pesticides can be broken down into compounds that are just as harmful or more harmful than the original pesticide. However, it must generally be expected that including factors for processing/preparation of the crops in the calculations will reduce the value of the intake. In future, final and well-tested calculations that include reduction factors and that are supported by toxicological data will become available, making it possible to compare the intake of the different pesticides.

Figure 6.4 Look here!
The total average intake of pesticide residues (denoted I) from different fruit and vegetables, when account is not taken of such factors as peeling and similar that reduce the residual content of pesticides. The vertical line gives the mean value for all pesticides (0.3% of ADI).

Figure 6.5 shows the main crops shown by the calculations to contribute to the daily intake of pesticides. It should be noted that account is not taken of peeling and other factors that reduce the residual content of pesticides. Since citrus fruit is eaten without the peel, and apples are eaten with the peel, the real relationship between these two intakes could be rather different from this figure. Secondly, it should be noted that the figure does not say anything about the health effects associated with the different pesticide residues in the different crops. For example, it could be misleading to report that apples, for example, contribute most to the intake of pesticides if toxicologists have shown that the pesticide residues that apples contain are less harmful than other pesticide residues.

Figure 6.5
The total average intake of pesticide residues (denoted I) from different fruit and vegetables in microgramme per day. In the calculation, account has not been taken of peeling and other factors that may reduce the residual content of pesticides in the food products.

(Figure text:
Citrusfrugter = Citrus fruit
Æbler = Apples
Vindruer = Grapes
Bær = Berries
Andre frugter = Other fruit
Kartofler = Potatoes
Andre grøntsager = Other vegetables
fra xx pesticider = from xx pesticides

Uncertainties and reservations

It must be remembered that the calculations are encumbered with relatively great uncertainty. That applies to the determination of the residues of the individual pesticides and to the dietary data used. At the same time, the residues used are based on a cautious evaluation, which means that the values must be assumed to lie above the real content. The result is thus an over-estimation of the intake of both Danish and imported products. There is also a big variation in the data used for intake of diet. This variation is due partly to uncertainty in the determination of the mean intake of the individual food products and partly to a variation as a consequence of different eating habits in the population. Taken overall, the intake of the individual food products shown in table 6.5 can vary by several hundred per cent from one person to another.

However, it must be stressed that a person with a relatively large consumption of one type of food will often have a smaller consumption of other types, so variations in the total pesticide intake will not necessarily vary to the same extent as the food consumption.

6.2.5 Determination of the effects of pesticides on the population

Absorption and excretion of pesticides

Uptake of pesticides takes place after oral intake and, to a lesser extent, through inhalation or via the skin. Absorption via the lungs is normally fast compared with the other exposure routes because of the thin alveolar membrane and the ample blood flow. Absorption via the skin is often slow, but for pesticides that are metabolised quickly in the liver, skin exposure can be the main exposure pathway. With respect to exposure of the foetus, it has been found in animal studies that certain pesticides – particularly organophosphates and carbamats – can pass the placenta (Salama et al. 1993).

Metabolism

Metabolism or biotransformation of the absorbed, biologically available dose (the internal dose) determines how large a part reaches the target organ. Almost all the chemical changes that pesticides undergo in the body are due to special enzymes (Hayes, Laws 1991). The first stage in the biotransformation normally takes place via microsomal enzymes, which catalyse an oxidation or reduction reaction. These enzymes include all the cytochromal P450-enzyme systems in the liver, which is the main organ for biotransformation of chemicals. The degradation products are normally less toxic and more easily eliminated than the pesticide itself. However, for some pesticides, an activation takes place, which can lead to the formation of more toxic metabolites.

Bioaccumulation

Some pesticides that accumulate in the body’s fatty tissue have very long half-lives. These so-called persistent substances are now banned. They include the chlorinated pesticides, such as DDT and dieldrin.

Authorisation and risk assessment of pesticides with a view to public health

According to an EU directive (91/414/EEC) on marketing of plant protection products, a risk assessment of health and environmental properties of the product must be carried out in connection with the authorisation procedure. The magnitude of the real and the potential exposure must be determined, both for the users and for the consumers. Users should be understood to mean spraying personnel, workers and others who are exposed to the product during and after application. The consumers are exposed to the pesticides through food and drinking water or, for example, via soil pollution.

Limit values

Allocating limit values for content in food products is part of this authorisation procedure. If a substance can be authorised for use, a maximum limit value, MRL (Maximum Residue Limit), is fixed for the maximum residual content of the pesticide or its degradation or transformation products in food products.

The limit value in food products is based partly on a toxicological assessment of the health risk in connection with intake of the pesticide in question, with an Acceptable Daily Intake (ADI) being set and partly on the residual concentrations of the pesticide found in vegetable food products after use of Good Agricultural Practice (GAP). GAP means the nationally authorised methods of use that are necessary under current conditions for effective control of pests. To arrive at the maximum limit value, the health aspect (ADI) is combined with the use of the pesticide in question (GAP). This is done by combining ADI with the Theoretical Maximum Daily Intake (TMDI), calculated by means of diet models and assuming that all crops for which the pesticide may be used contain the maximum permissible amount of the pesticide. TMDI must lie below ADI for a limit value to be set. In Denmark, it is the Danish Veterinary & Food Administration that sets the limit values for food products.

Authorisation of pesticides

In connection with the authorisation procedure, not only the pesticide’s physical/chemical properties and its stability and degradation in nature must be tested, but also its toxic properties. There are internationally agreed guidelines for performance of the toxicological tests that are included in the risk assessment of pesticides. Both the active ingredient and the product must thus be tested for acute effects, and the active ingredient must be tested for chronic effects through repeated, long-term exposure (carcinogenicity, mutagenicity and effects on reproduction, including malformation of the offspring). In addition, tests are required concerning the substance’s absorption, metabolism, accumulation and elimination, together with any effect on enzymes and other biochemical parameters. In the toxicological testing of chemicals, account must be taken of the substance’s purity, stability in the exposure set-up and reproducibility in a possibly repeated test.

In the nature of things, most of the testing of the health effects of pesticides is done on animals (usually mice and rats) rather than people. There can be differences between test animals and humans with respect to metabolism, and far higher doses are normally used in the animal tests than the general population could be exposed to.

Acute reference dose

For certain pesticides, however, there is only a rather narrow margin between the doses that are acutely toxic and the doses that have long-term effects. For such pesticides, an acute reference dose is set in addition to the usual ADI as a protection against acute toxic effects. Readers are also referred to section 8.3 on the precautionary principle.

Measurement of the potential exposure

The potential human exposure to pesticides can be investigated by direct measurements in food products, drinking water or the environment. The most accurate data for exposure are presented in case reports, but only a group designation is usually used (e.g. insecticide, fungicide or herbicide) when describing the exposure, and the amount used is self-reported. In several studies of children and pesticides, the length of time the parents have been engaged in an occupation with potential pesticide exposure, for example, and the frequency and use of pesticides are used as indirect measures of the degree of exposure. Other studies have estimated the magnitude of the exposure by combining types of crops in an area with information about the use of pesticides that are specific to the crops in question. The rougher the measure used for classification of the type and extent of the exposure, the more likely it is that any real increased risk of health effects from specific pesticides will not be discovered.

The indicators can be too rough

In many of the studies, an assessment of effects is also often based on rough indicators. An exact description of the effects is extremely important but is often problematical – as in the case, for instance, of neurotoxic effects of pesticides, which can be very difficult to reveal and quantify. Assessments are frequently based on comparisons of mean values, whereby particularly deviant small groups can be overlooked. For cancer, tissue tests may be necessary in order to make specific cancer diagnoses. Sub-types of some diseases, e.g. leukaemia, can have different causes, and mixing these sub-types can blur a possible relationship with a given exposure. Owing to poor precision in assessment of both exposure and effect in the traditional epidemiology, incorrect classification is likely, which results in these methods having a low sensitivity, see below.

Epidemiological studies

Information on relationships between exposure to pesticides or other environmental factors and the occurrence of disease in humans is often obtained by means of epidemiological studies (also called population studies). Great caution must be exercised in interpreting the results of epidemiological studies because of a number of sources of error and inherent weaknesses in the different epidemiological models. It must be firmly stated from the start that epidemiological studies - in the classic sense – are not enough on their own to establish causal relationships.

The models comprise: 1) descriptive epidemiological studies (population-based) and 2) analytical epidemiological studies (individual-based).

Cross-sectional studies (descriptive studies)

In descriptive epidemiological studies, the lifestyle or conditions of life of groups (e.g. whole countries’ populations) are typically registered as causal factors. For example, the consumption of a given pesticide in different population groups can be compared with statistical data on morbidity and mortality in these population groups, or changes in the consumption of pesticides and morbidity or mortality over time in the same population group. An obvious drawback with descriptive studies is that it is not possible to make adjustment for known risk factors at individual level (e.g. smoking, obesity or alcohol consumption). In descriptive studies one can find certain patterns between exposure and disease. For example, a high consumption of pesticides and thus probably exposure to them in large, limited population groups (e.g. farmers) can be linked to an increased occurrence of cancer. However, such a relationship says nothing at all about causal relationships, i.e. that a high exposure to a pesticide cannot, from the above data, be said to be the cause of cancer but can give rise to a suspicion concerning causal relationships, which must then be examined more closely.

Analytical studies: case control studies

Case control studies are the most common type of analytical epidemiological studies. In these, one compares causal factors, e.g. lifestyle, in individuals with a given disease or other factor that one wishes to investigate, with healthy persons. The individual in the control group should be chosen at random, and with the same inclusion and exclusion criteria, from the same population base as cases. Case control studies are retrospective. Memory therefore plays a vital role. It is also the case that precisely the memory of a sick person is often seriously affected by the disease and can give rise to systematic errors (bias). It is also often uncertain which period of the sick person’s life is most relevant for the study. The longer the time elapsing between a harmful effect and the occurrence of recognised disease, the greater this problem becomes. A study that compares the pesticide intake through diet in women that have borne a child with a congenital defect with the intake in women that have born healthy children is an example of a case control study that is encumbered with these possibilities of error.

Analytical studies: cohorts

Another frequently used analytical epidemiological study – the prospective cohort study – has fewer weaknesses than case control studies. In the prospective cohort study, relevant information is gathered from healthy persons, who are then monitored for a period – often a number of years – during which the morbidity of the cohort is recorded. This means that when a disease occurs one has the possibility of going back and finding the factors that characterised persons that later became ill for comparison with persons that did not do so. In cohort studies one also has the possibility of extracting and storing biological material for later analysis. For example, analyses of blood samples, fatty tissue samples and milk from nursing mothers provide information about the exposure to pesticides.

Cohort studies have the major drawback that they require many participants because the incidence (number of new cases per year) of ordinary diseases (e.g. cancer) is relatively low. Another factor is that it can take a long time for a disease to develop. In the case of cancer, 20 years or more can elapse from the initiation phase until clinical diagnosis (see later). These factors mean that cohort studies are considerably more costly and more difficult to handle than case control studies.

However, in the case of cohort studies as well, it is not possible to determine whether there is a causal relationship. Persons with a high exposure to a pesticide probably differ in other ways from persons characterised by a low exposure. In studies, one often adjusts statistically for factors that can confound the result, e.g. socioeconomic status, smoking habits, alcohol consumption, physical activity, use of hormones, etc. These factors are called "confounders" or confounding factors. However, it is important to be aware that differences in, say, lifestyle and conditions of life cover a very large number of factors and too little is known about their effect on the development of disease.

Other sources of error

Lack of big differences in exposure in the population group studied can be a problem in studies of the effect of pesticides on public health. The sources of error mentioned above are just a few of the inherent sources of error in epidemiological studies. The uncertainty concerning, for example, intake of pesticides is a particular problem when there is only a small variation in intake in the population group studied. While the true difference in risk of, say, cancer between a lower and upper level of a narrow intake interval is probably small, the observed difference decreases with increasing uncertainty of the analytical measurements. A negative result, i.e. a lack of relationship between exposure to a pesticide and the risk of disease can therefore not be taken to mean that such a relationship does not exist.

For a more detailed discussion of confounding and other sources of error in epidemiological intake/exposure studies, see Tarasuk, Brooker 1997.

No conclusive evidence

It cannot be proven on the basis of epidemiological studies that pesticides, in the quantities to which the general population is exposed, for example, via diet, are harmful to health. Conversely, one can never completely prove scientifically that a pesticide cannot cause a health risk. One can, however, show, with greater or lesser (un)certainty, the probability or lack of probability of a health risk. This applies to all scientific work, including tests carried out on animals.

Low-dose exposure implies lengthy, continued or occasional exposure to low levels of pesticides. With this form of exposure the body slowly accumulates a quantity of one or more pesticides that is sufficient to produce undesirable effects in the body. Unlike the health effects of occupational exposure to pesticides (see section 6.1), there are very few epidemiological studies of the relationship between exposure of the general population and the risk of developing a disease. In the following, we sum up the findings from existing studies of the effect of pesticides on reproduction, the hormonal system, cancer, the nervous system and the immune system. For further amplification of the existing knowledge, readers are referred to Skadhauge (1998).

Reproductive toxicity

Toxic effects on reproduction can take the form of difficulty in conceiving (reduced fertility or unrecognised early miscarriage), infertility, miscarriage and stillbirth (Smith et al. 1997). In women, the effects can also manifest themselves in the form of paramenia (as a consequence of changes in the neuro-endocrine function in the hypothalamus, the pituitary gland and the ovaries) and, in men, in the form of a low sperm count or changes in the mobility or appearance of the sperm.

In animal tests, several pesticides have shown toxic effects on reproduction (Traina et al. 1994). Particular attention has been paid to pesticides with hormone-like() effects, but, in animal tests, a number of other chemical substances have proved able to affect the reproductive system, resulting in, for example, reduced quality of semen and infertility without it being possible to point directly to an oestrogenous effect. As stated earlier, several studies of individuals with a high occupational exposure have shown negative effects on the reproductive system and particularly on fertility. However, in the case of the general population, the data are much less certain.

In 1995, DEPA published a report that concluded that sperm quality in otherwise healthy men had been falling since the end of the 1930s (DEPA 1995a). In the same period, a marked increase in testicular cancer was recorded, particularly in younger men. An increased frequency of certain deformities of the sexual organs of boys also seemed to have occurred. However, there was insufficient documentation to determine whether (early) exposure to pesticides could have been a cause of these effects on the reproductive system. In the case of most chemical pollutants in our environment, it is unknown 1) whether they have an oestrogenous effect or not, 2) how big their effect is, individually or combined, and 3) the actual magnitude of the exposure (Editorial 1995).

Cancer

For a number of pesticides previously in use there is evidence of carcinogenicity in animals, whereas arsenic and mixtures containing arsenic are the only pesticides classified by IARC (International Agency for Research on Cancer) for which there is sufficient evidence of carcinogenicity in humans, based on an increased risk of lung cancer and skin cancer.

There are currently some few pesticides on the market that are classified as carcinogenic in group 3 (Carc 3), see section 8.1 and Lindhart et al. (1998).

Mechanism behind the development of cancer

The carcinogenic effect can be exercised in different ways. Some pesticides affect the cells’ DNA and, in the worst scenario, induce precisely the changes that enable the cells to develop malignant properties (initiators). Other pesticides can cause cancer to develop by stimulating initiated cells to divide further (promotors). In relation to pesticides, attention has been paid particularly to their hormone-like effect, since oestrogens, for example, can promote hormone-dependent initiated cells (e.g. mammary gland cells). However, pesticides may also affect the development of cancer in other and more indirect ways, for example by inhibiting specific parts of the immune system.

Cancer cannot be predicted

One big difficulty in investigating a relationship between exposure to an environmental factor and the risk of cancer is that cancer often takes a very long time to develop – up to 20 years and perhaps even longer. Despite intensive research, valid biomarkers for early stages of cancer have not been identified with certainty.

Findings in epidemiological studies

There are only a few epidemiological studies of the relationship between the risk of cancer and exposure to pesticides in the general population. Most of the studies have focused on the relationship between the level of organochlorine compounds in the blood and the risk of breast cancer. In 1993, a study from New York aroused some interest (Wolf et al. 1993). In this study, an approximately 4 times higher frequency of breast cancer was found in women with a high blood concentration of DDE and PCB. Dieldrin was not measured. In 1994, a similar study was carried out in California, and there, the relationship was not nearly as strong, even though the exposure to DDT was considerably greater (Krieger et al. 1994). A third major American study was unable to establish any relationship between the blood content of organochlorine compounds and the risk of breast cancer. However, the follow-up period was short, and problems with confounder correction resulted in statistical weakness (Hunter et al. 1997). Lastly, a recently published Danish study showed a statistically significant dose-related correlation between the blood content of dieldrin and the risk of breast cancer (Høyer et al. 1998). The other relationships between risk of cancer and blood content of beta-hexachlorocyclohexane, DDT and PCB were not statistically significant.

Relationship between pesticides and cancer not documented

As the results of the above-mentioned epidemiological studies indicate, there are data that show that exposure to organochlorine compounds can be connected with an increased risk of breast cancer, although without sufficient, clear grounds to deduce the significance for the general population. A scientific panel appointed by the National Cancer Institute of Canada came to a similar conclusion in a review of studies of mainly occupational exposure and risk of cancer (Ritter 1997). The panel concluded that there was insufficient scientific evidence that pesticides contributed significantly to total cancer mortality.

Hormonal toxicity

It is well documented that a number of persistent chemicals can affect the endocrine system by affecting the hormones in the body that are responsible for maintaining homeostasis and regulating the developmental processes (Gray, Ostby 1998; Kavlock et al. 1996; Tilson, Kavlock 1997; Porter et al. 1993). Endocrine disruptors have been broadly defined as exogenous substances that disrupt production, release, transport, metabolism, sorption, action or elimination of the body’s natural hormones (Kavlock et al. 1996; Tilson, Kavlock 1997). Some of these substances can induce the cytochromal P450 systems in the liver. This can affect the metabolism of steroid hormones (including sex hormones) and thyroid hormones, whereby secondary effects can be induced in hormone-dependent organs.

However, too little is known about endocrine disruptors, and the designation has in some cases been used where the biological relationship is far from clarified. The suspicion about the very persistent substances DDT, DDE and PCB is warranted in view of the effects these substances have on wild animals (e.g. seals, otters and alligators). For the endocrine disruptors, too, there are problems in extrapolating from studies of test animals with relatively high exposures to the relatively low exposures that occur in the environment and the very small amounts that humans are exposed to. These endocrine disruptions cannot be demonstrated in present test systems, but work is going on under OECD to improve the test programme for new chemicals, so that animal tests are better able to document such effects.

Developmental toxicity

Developmental effects are effects that occur in the individual during development as a result of exposure prior to conception, in the foetal existence or postnatally until the individual is fully developed. The result of the developmental disturbances can be death, structural abnormalities, inhibited growth or functional disturbances. However, it is often difficult to determine whether miscarriage or stillbirth is due to reproductive or developmental effects.

There are too few studies on the relationship between exposure to pesticides and developmental effects. This also applies to the potential of pesticides and fertilisers in contaminated groundwater to produce reproductive and developmental toxic effects in humans.

Neurotoxicity

The term neurotoxicity is used to describe a substance’s potential to change the structure or function of the nervous system, defined by neurochemical, neuropathological organ changes, behaviour and specific psychological processes such as sense perception, learning and memory. Many of these disturbances can be directly measured with neurochemical, neurophysiological and neuropathological methods, whereas others can only be interpreted by looking at the behaviour. The action mechanism of many groups of pesticides – particularly the insecticides – the intended target of which is the peripheral nervous system, leads directly to consideration of their potential neurotoxicity, whereas the neurotoxic potential of other groups is less clear. Studies of neurotoxic effects are difficult to perform and interpret, for which reason there are very few conclusive studies.

Experimental studies have shown that the nervous system is more vulnerable to exposure to chemical substances during development than when it is fully developed. Animal tests have shown that neurotoxicity as a consequence of low-dose exposure to such pesticides as organophosphates, pyrethroids, DDT and paraquat during development of the brain can lead to irreversible changes in the adult brain and induce behavioural and cholinergic changes in the adult animal, whereas exposure of an adult to the same substance has little or no effect (Eriksson 1997; Williams et al. 1997). Exposure to chemical substances during development of the nervous system can vary, both qualitatively and quantitatively, depending on the phase of development of the nervous system (non-linear dose-response curve). It can therefore be difficult to assess the effects of long term low-dose exposure on the basis of short-term studies. In addition, the development of the nervous system depends on the endocrine systems, which are responsible for sexual development and growth and are closely associated with the presence of circulating thyroid hormones. Moderate to major changes in the concentration of thyroid hormone during development result in motoric dysfunction, cognitive defects and other neurological abnormalities (Porterfield, Hendrich 1993). As stated, it can be difficult to assess neurotoxic effects. The National Research Council (1993) found that the present strategy for testing toxicity was inadequate for assessing toxicity to several organ systems, including neurological development processes.

In several epidemiological studies, Parkinson’s disease has been linked to pesticide use (including Semchuk et al. 1992). The main suspect was paraquat, the chemical structure of which is somewhat similar to a chemical (MPTP), which can induce parkinsonian symptoms in laboratory animals. There are no studies specifically elucidating the effect of paraquat.

Neurotoxicity in children

There are very few studies of children and neurotoxic effects as a consequence of exposure to pesticides, and clear conclusions cannot be drawn from them. There seems to be insufficient documentation of neurotoxic effects in the fully developed nervous system as a consequence of low-dose exposure to pesticides. However, experimental data provide a basis for assuming that the nervous system is vulnerable to pesticide exposure during development, partly due to disruption of endocrine systems, which are of importance for normal development of the nervous system.

Immunotoxicity

The toxicological tests required as a minimum before a plant protection product can be authorised can to some extent show whether a chemical substance has a potential immunotoxic effect when the organism is exposed to it. However, the test methods have some constraints, particularly with respect to sensitivity, i.e. their ability to show any weak immunotoxic effects. Generally speaking, chemical substances with a clear immunotoxic effect, including pesticides, will be identified in the required toxicological tests, whereas there can be a problem with substances with a medium to weak effect.

Immunotoxic effects of chemical substances, including pesticides, can occur as a consequence of the immune system being either weakened or strengthened, or if the immune system responds to the chemical substance as an antigen, which can lead to allergy or autoimmunity.

Several animal studies have shown immunological changes after acute or chronic exposure to pesticides (Banerjee et al. 1996; WHO 1996). In many cases, the results from these studies are difficult to interpret, and some of the studies could indicate that the immunotoxic effect can be a consequence of another, systemic, toxic effect. However, the tests show that pesticides, as well as other chemical substances, can have a potential effect on the organism’s immune system (WHO 1996).

Animal tests using rodents indicate that the immune system in the developing organism is more vulnerable than the fully developed organism. Experimental and clinical data from the working environment have also shown that chemical substances, including certain pesticides (chlornitrobenzene, captan og organophosphates) can induce contact hypersensitivity (type IV reaction) in test animals and humans (Cronin 1980).

The epidemiological evidence of immunotoxic effects as a consequence of exposure to pesticides is sparse. Immediate reactions (type I reaction) are thus only described in a few case reports that have been difficult to confirm (Baker, Wilkinson 1990).

Data from studies of chemical substances, including certain pesticides, show that these can affect the immune system under special experimental conditions. The tests on which the toxicological assessment is usually based can identify the pesticides that have a potential serious effect on the immune system. With the exception of contact hypersensitivity as a consequence of occupational exposure to pesticides, there are insufficient data to conclude that exposure to pesticides in the environment in normal circumstances will cause damage to health in the general population as a consequence of an effect on the immune system (Baker, Wilkinson 1990; Banerjee et al. 1996; WHO 1996).

6.2.7 Conclusions

There is insufficient epidemiological evidence to either prove or disprove a relationship between health effects and long-term low-dose exposure to pesticides. There are several reasons why there are considerable difficulties in gaining more certain knowledge about the effect of pesticides on human health. The epidemiological studies concerning the effects on humans of exposure to pesticides are characterised by imprecise measures for both exposure and effect, a relatively short follow-up period and lack of control of confounding factors. Due to limited group sizes, data are often collected in large groups, which further reduces the sensitivity. Owing to low exposure contrast and many confounding factors, there are often problems in proving any effects.

The National Board of Health has carried out a study of hospital admissions, outpatient attendance and deaths in Denmark caused by pesticides in the period 1987-1996, using data from the National Registry of Patients and the Cause of Death Registry. However, the admission figures from the period before 1994 are encumbered with incorrect diagnostic codes (Skadhauge 1998). In the period 1994-1996, there were 88 admissions in all. Of these, 48% were children under the age of 11 years and 39% were women. With respect to outpatients, the first countrywide records are from 1995. 127 cases were recorded in 1951 and 51 in 1996. However, some of the codings in 1995 could be incorrect diagnostic codes. A manual examination of death certificates revealed 48 deaths, 45 of which were suicide and three accidents in children. For the admissions and outpatient cases, it is not possible to determine whether the poisoning occurred in connection with work in farming. The three child deaths all occurred in rural communities. Two of the cases were due to unlawful storage in unmarked packaging.

The Poisons Information Office of the Occupational and Environmental Medicine Clinic at Bispebjerg Hospital received 163 calls about pesticide poisoning in 1997. 47% concerned children aged 0-14 years. The calls covered the following groups of substances: