General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (2/05) Page 1 of 25

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

A SID 5A form must be completed where a project is paid on a monthly basis or against quarterly invoices. No SID 5A is required where payments are made at milestone points. When a SID 5A is required, no SID 5 form will be accepted without the accompanying SID 5A.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code PS2209

2. Project title

The environmental impact of sprayer decontamination

3. Contractororganisation(s)

Central Science LaboratorySand HuttonYorkYO41 1LZ          

54. Total Defra project costs £ 119,994

5. Project: start date................ 01 July 2004

end date................. 30 September 2005

SID 5 (2/05) Page 2 of 25

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.External residues on agricultural sprayers can be detrimental to the longevity of the machinery, cause injury to subsequently treated crops, and there could be potential health implications for persons coming into contact with contaminated sprayers; sprayers should therefore be cleaned after use. However, there are concerns that the residues removed could adversely impact on the environment.

The aim of this study was to quantify residues associated with cleaning sprayers, and to explore practical methods by which the operator can minimise any potential environmental risks. The work was carried out in two phases; ‘laboratory’ based studies were conducted under controlled conditions to examine factors that may influence the efficiency of cleaning methods, and field visits were undertaken to quantify pesticide loadings contained in washings from working sprayers.

Representative sprayer surfaces were used to investigate the suitability of the tracers, disulphine blue, tartrazine, fluorescein, and waxoline red. The experiments also investigated the performance of different classes of ‘stickers’ (adjuvants) including Bond, Sprayfast, Phase II, Activator 90, and Designer. The initial experimental method used 14L pure water tanks, the spray solution was applied by an airbrush and three replicates were used. It became apparent that highly water soluble compounds are largely ineffective tracers for the investigation of external sprayer cleaning, even when mixed with ‘stickers,’ and identifying a suitable tracer was far more difficult than envisaged. It was proposed that this was due to the inherent solubility of the tracers in conjunction with the relatively high volumes of water used rather than any characteristic of the stickers. The oil-soluble dye, waxoline red was unsuitable due to extraction and analysis difficulties. A white, water-based paint emulsion was ultimately identified as a suitable tracer. This was applied to a 0.7m2 area on one side of a 300L spray tank surface using a paint brush, and the mass applied was calculated gravimetrically by difference. The mass removed was calculated by multiplying the volume of the washings by the concentration of the white paint. These surfaces were used to investigate the cleaning efficiency of a hot water high pressure washer (c. 100 bar), a cold water high pressure washer (c. 100 bar), a low pressure onboard washer (c. 10 bar) (as used onboard sprayers for in-field The low-pressure onboard washer was relatively inefficient at removing residues (30% removed) compared to the high-pressure washer (70- 80% removed), which had comparable performances regardless of water temperature; the brush proved equally as efficient as the high-pressure washer (80%). However, the flow rate to the brush was relatively high (c. 10 L min-1), thus it may be advantageous for manufacturers to consider lowering the water rate to optimise the use of the limited water supply for in-field cleaning.

The influence of distance between the lance and target (0.25, 0.5, 1 and 1.5m), and the angle (45° and

SID 5 (2/05) Page 3 of 25

90°) of water impact was also investigated using the cold water high pressure washer. There was a decrease in cleaning efficiency with increasing distance from lance to target, and this was most noticeable at a distance greater than 1 m. There was also a reduction in cleaning efficiency when the angle of impact was 45° compared to 90°. There was a marked decrease in cleaning efficiency, from c. 70% to 30% or less, at a distance of greater than 0.5 m when the angle of impact was 45°.

To quantify the residues contained in washings from working sprayers, the spray operator cleaned the sprayer in four discrete sections: the boom and back of the sprayer, the rear wheels, the spray tank, and the cab/tractor. A tarpaulin was placed beneath each area to collect the washings; a different tarpaulin was used for each of the discrete sections. A litre sample of the washings was taken and this was analysed for azxoystrobin, carbendazim, chlorothalonil, flusilazole, isoproturon, tebuconazole, pirimicarb and pendimethalin. Mud samples were also taken when it was liberally coated on the sprayer. The flow rate of the washer was measured, and the volume of water used during cleaning was calculated from the time taken to clean multiplied by the flow rate. The mass of residue removed from each discrete area (concentration x water volume) was used in the data analysis.

The boom and the back of the sprayer contributed 80% of the total pesticide loading; mud may also be a significant contributor to pesticides in farmyards. Sixty percent of the samples contained pesticide residues in excess of 0.1 µg L-1 with 48% of these emanating from the boom and tank, but this was compound dependent. The majority (80%) of pirimicarb, flusilazole, azxoystrobin and carbendazim concentrations were below 0.1 µg L-1, and the higher concentrations could largely be attributed to washing the boom alone. On the whole, predicted environmental concentrations arising from sprayer washing were at least one order of magnitude lower than PECs generated from the FOCUS surface water model (Step 1). The quantity of residues removed during cleaning represented a very small fraction of the typical application rate thus cleaning the sprayer in the field is unlikely to impact adversely on the environment.

A high pressure washer or hose with brush may provide the most practical method for cleaning the external surfaces of sprayers. However, if a pressure washer is used, care must be taken to ensure that the lance is no more than 0.5 m from the target surface, and that the angle of impact is as close to 90° as possible. In practice, this would mean that the boom should be washed when unfolded and lowered. This would have implications for washing down sprayers over restricted catchment areas such as those typically used for drive-over biobeds, and an offset system may be more suitable, as long as the bunded area was sufficiently large. Whilst the environmental risk arising from cleaning the sprayer in the farmyard may be low, any risk should be negated if the sprayer is cleaned in the field. If there is only a limited water supply available for use in the field it may be advantageous to focus this on the boom and back of the sprayer, the spray tank and muddy areas.

Both the potential health and environmental risks associated with sprayer cleaning should be considered when providing guidance to spray operators, and the results of the current study will be amalgamated with the findings of a recent study by HSE investigating occupational exposure to pesticides during cleaning to provide unified guidance to spray operators on the external cleaning of sprayers.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met;

SID 5 (2/05) Page 4 of 25

details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

1 INTRODUCTIONDuring the application of pesticides, residues are deposited on the external surfaces of the sprayer and doses exceeding 1 mg m-2 are not uncommon (Ramwell et al., 2004). These residues can be detrimental to the longevity of the machinery, and/or cause injury to subsequently treated crops. More significantly, external residues could have potential health implications for persons coming into contact with the sprayer (Ramwell et al., 2005a). Sprayers should therefore be cleaned after use and a common cleaning method is the pressure washer (Ramwell et al., 2002). However, cleaning the sprayer in the farmyard has been identified as a potentially significant source of pesticides detected in surface waters (Fischer et al., 1998). A study in the UK found that 0.0007% of the product applied in the field (excluding the contribution from mud) was available to be washed off during cleaning (Higginbotham et al., 1999). This compares to a similar previous study where 0.04% of the product applied in the field was potentially available to be removed (Fogg, 1999). Both these studies investigated a single compound (isoproturon) after a single spraying event, yet the results were two orders of magnitude different emphasising the caution required in extrapolating the results from such small data sets. Indeed, a larger scale study investigating

SID 5 (2/05) Page 5 of 25

residues of thirteen commonly used pesticides on the external surfaces of different sprayer types highlighted the variation in residue levels that can occur both between and within compounds and sprayers (Ramwell et al., 2004), thus the potential for contamination arising from sprayer cleaning may also vary widely.

There have been a number of developments with regard to sprayer cleaning over recent years (Balsari et al., 2002, Holst et al, 2001). In the UK, the Voluntary Initiative has been a significant driver of this and there is the need to have a greater accountability of the route of pesticides into the environment. However, there are some concerns that removing external sprayer residues could have an adverse impact on the environment, but there are no data to either negate, or support this theory. In practice, different parts of the sprayer will be more contaminated than others, notably the boom and the back of the sprayer. It is possible therefore that any environmental risk associated with cleaning a sprayer may arise from washing down specific parts of the sprayer.

Many sprayers now carry onboard cleaning kits that allow the machines to be washed down in the field thus keeping pesticides in the field of use, and avoiding the generation of any waste that would require groundwater authorisation for disposal. Cleaning the sprayer in the field also avoids any potential contamination of the land across which the sprayer must travel, for example via mud deposits contaminated with pesticides on the road. However, there are potential penalties to the operator associated with cleaning the external surfaces of the sprayer in the field. First, sufficient water must be carried on the sprayer to clean both the internal and external surfaces. The volume of water should be kept to a minimum to avoid costs involved with manufacturing large clean water tanks and transporting them around the field during spraying, the extra weight of which could add significant costs to fuel bills. Second, infield cleaning kits run at a lower pressure than industrial pressure washers used on many farms, thus infield cleaning kits may not be as efficient as washers available in the farmyard, and their use may be more time consuming. Knowledge on the pesticide loadings from discrete areas of the sprayer could assist in optimising the use of a limited water supply, enhancing the practicality of cleaning in the field and ultimately reducing the environmental impact of pesticide use.

The aim of this study was to quantify residues associated with cleaning sprayers and to explore practical methods by which the operator can minimise any potential environmental risks.

The objectives of the study were to:

1. Conduct preliminary investigations under controlled conditions to examine factors that may influence the efficiency of cleaning methods and to establish typical proportions that different areas of the sprayer may contribute to the total pesticide loading of the washings; the methods used and the results of this phase of the work are presented in Section 2.

2. Sample working sprayers to establish typical, realistic pesticide loadings from washing sprayers; the methods used and the results of this phase of the work are presented in Section 3.

3. To propose mitigating strategies as appropriate in relation to potential environmental risks; this work is presented in Section 3.

2 PRELIMINARY INVESTIGATIONSTests were conducted under controlled conditions to explore factors influencing the removal of external residues. There were three main work areas: identifying a suitable tracer, examining influential factors, and quantifying losses from discrete areas of the sprayer under controlled conditions.

2.1 Identification of a suitable tracerThe use of tracers is generally accepted as a practical alternative to using pesticides in studies involving sprayers. The advantages include significantly cheaper analysis, no exposure of the researchers to pesticides, and no exposure to the environment. The tracers used are commonly highly water soluble and they may be mixed with other compounds, such as adjuvants, to more accurately mimic pesticide behaviour. Typical tracers include tartrazine, copperoxychloride, disulphine blue, and fluorescent dyes (e.g. Balsari et al., 2002; Holst et al., 2001; Barber & Parkin, 2003).

For the purposes of this study, a suitable tracer would be one that: was not excessively time consuming to analyse, could be analysed at concentrations relevant to the study, posed minimal risk to both the environment and the researcher, and adhered sufficiently to the test surface to provide some resistance to being removed during cleaning. Numerous experiments were conducted to identify a suitable tracer, and this proved far more

SID 5 (2/05) Page 6 of 25

difficult than was originally envisaged. The results and discussion from these experiments are combined below to provide a succinct overview of the work.

2.1.1 Small scale experimental apparatusPure water tanks (14L) were used as the representative test surface with three replicates per test. The spray solution was applied using an airbrush fitted with a glass vial reservoir (1.8 ml). The target application rate was 5 mg of the tracer for each surface; this mass is comparable to the mean pesticide mass on the external surfaces of spray tanks and mudguards of field sprayers (Ramwell et al., 2004). Surfaces were left to air dry for 24 hours prior to washing. Water from the lance of a cold water, light industrial pressure washer (Nilfisk Alto) was passed over the tank three times for a duration of 5 seconds; this equates to a representative cleaning time of approximately 20 minutes for an entire sprayer. The washings were collected in a 60L vessel placed beneath the pure water tank and a protective sheet was used to ensure adjacent tanks were not inadvertently splashed. The experimental apparatus and method of cleaning was identical to a previous study examining the decontamination of sprayers in relation to occupational exposure (Ramwell et al., 2005a), thus enabling direct comparison of the results where appropriate; for reference, approximately 40% of pesticides applied could be removed by cold pressure washing in the aforementioned study although there was variation between compounds (range 25% to 70%). The apparatus was used to investigate the behaviour of a number of tracers and sticker compounds. The specifics of each test and the results are presented below.

Disulphine blue

Disulphine blue was applied to the 14L tanks using distilled water as the solvent. Removal rates were greater than 95%, thus the experiment was terminated.

Disulphine blue and tartrazine, plus cellulose

A stock solution of viscous dye solution was prepared gravimetrically to BS standards, i.e. 1% xanthan gum, 0.3% cellulose and 0.1% dye. This Standard is for the emulation of a concentrated pesticide product. A representative dilution was prepared using the assumptions of the product being applied at 5 L ha -1 with 200 L ha-1 of water. This gave a dilution factor of 40. Each dye plus viscous solution was prepared and applied separately.

Removal rates were greater than 95% for both dyes, the viscous solution was difficult and time consuming to prepare, and it can go off, thus this preparation was excluded from further investigation.

There was no significant difference in removal rates for the disulphine blue and tartrazine. For the purposes of this study, disulphine blue was chosen for further investigations, primarily because the sprayer tanks were yellow, thus it was very difficult to make any visual observations on the behaviour of tartrazine during and after washing.

Disulphine Blue + adjuvants

Adjuvants are compounds devoid of pesticidal activity that, when combined with a formulated pesticide, enhance some property of the active ingredient. There are several types of adjuvants, one of which is a ‘sticker’. These compounds act to enhance the adhesion of the pesticide to the target. It was reasoned that sticker adjuvants could reduce the ease with which tracers are removed from a sprayer surface during washing. Several adjuvants were investigated to ensure the range in possible chemical classes was considered (Table 1).

Table 1 Adjuvants investigated

Product Constituent compound(s)Sprayfast PinoleneActivator 90 750g/L alkylphenylhydroxypolyoxyethylene

150 g/L natural fatty acidsBond 450 g/L synthetic latex

100g/L alkylphenylhydroxypolyoxyethyleneDesigner 50% styrene-butadine copolymer

8.44% polyalkylene oxide modified heptamethyl trisiloxonePhase II methylated rapeseed oil

The results of the test indicated that there was a tendency for adjuvants to increase the adhesion of disulphine blue to sprayer surfaces (Figure 1); this effect was most noticeable for Bond. However, the variation between the replicates was too large to have great confidence in the findings. It was possible that raising the number of replicates could identify whether the effect of Bond was real or not, and/or using larger test surfaces could reduce natural variation between replicates and enhance any effect of the adjuvants. Although the removal rates of Bond

SID 5 (2/05) Page 7 of 25

were still relatively high (70%) it was postulated that a longer drying time could enhance the adhesive effect, thus further experiments were planned.

Figure 1 The effect of adjuvants on the removal of dye showing the mean and ± 1se

Waxoline red

Waxoline red is an oil soluble dye that should provide more resistance to being removed from a sprayer surface by water. Experiments were conducted to investigate this theory. The spray solution could be applied to the surface evenly using cooking oil as a diluent. The dye resisted removal when cleaned, but difficulties were encountered at the analysis stage. The concentrations of dye in the water were low and the oil was visible as a poorly distributed emulsion. Attempts to extract the dye using ethyl acetate and acetone were not successful. Chloroform may have been a more efficient solvent for extraction, but given the number of samples forecasted and the hazardous nature of this solvent, the extraction of the dye with chloroform would not be resource-efficient, and it was considered too impractical.

2.1.2 Medium scale experimental apparatusIn an attempt to reduce natural variation that may be attributable to scale, subsequent experiments were conducted with 300L spray tanks. This represented the maximum size of spray tank that could be used practically when considering manoeuvrability for research purposes. Initial attempts at using the airbrush to apply the spray solution were hampered by the low ambient temperature causing the airbrush to intermittently malfunction. Consequently, a standard household mister was used to apply the spray solutions. An area (0.7 m2) on one side of the tank was demarcated, inside which the spray was applied. The quantity of spray applied was calculated gravimetrically by difference before and after spraying. The tanks were left to air dry for a specified time prior to washing at which point it was transferred to a walled, drainer which directed all the washings into a single collection point.

Disulphine Blue + Bond

To investigate the suitability of Bond in increasing the adhesion of disulphine blue to sprayer surfaces, two dilutions were considered initially: 10% and 1%; this compares to the field application rate of 0.14%. The tanks were left to dry under a warm flow of air for 1 hour prior to washing with a cold water pressure washer with the lance being passed across the tank in a horizontal swath four times.

It was observed that there was a tendency for the spray solution to run down the tank during application rather than remaining evenly covered across the surface; it is possible that the Bond increases the size of the droplets produced, reducing both the drying ability of the spray, and the homogeneity of coverage, thus the spray solution can gather momentum to flow across the surface. This effect was more noticeable for the 10% Bond solution than the 1%.

There was visibly more dye remaining on the sprayer surface when the higher rate of Bond was used, particularly where the spray solution had ‘run’. It was also observed that there was less re-distribution of the dye to outside the treated area during cleaning for the higher concentration of Bond.

Although these findings were positive, there were concerns with regard to the application method and the fact that the spray solution could drip. Further tests were conducted to explore different diluents that would evaporate more easily than water and therefore reduce the drying time.

SID 5 (2/05) Page 8 of 25

Nortron as a diluent

Nortron is a xlyene-based ‘blank’ formulation that is used with active ingredients to create a formulated product. Three spray solutions were investigated to assess the effectiveness of Nortron as a diluent, namely :1. Water + disulphine blue

2. Nortron + disulphine blue

3. Notron + 10% Bond + disulphine blue

Each solution was sprayed onto the single side of a 300L tank, left to dry for 1 hour, and washed off using a knapsack fitted with a 4-10-12 nozzle. The ‘washing’ regime consisted of three individual passes of the lance against the tank. Observations were as follows:

Water + disulphine blue1st pass – the dye became more visible (presumably the dye was wetted up) and some ran across the surface, but little was removed in total.2nd pass – a large proportion of the dye was removed from the treated area, with some of this being re-distributed to the untreated areas of the tank rather than being washed off.3rd pass – the treated area was visibly clean.

Nortron + disulphine blue1st pass – the dye became more visible and there was some movement of dye within the treated area.2nd pass – More dye remained visible in the treated area than with water as a diluent; redistribution of the dye around the tank occurred.3rd pass – the treated area was visibly clean.

Nortron + disulphine blue + 10% Bond1st pass – the dye became more visible and there was some movement of dye within the treated area.2nd pass – More dye remained visible in the treated area than with water or Nortron as a diluent. There was some redistribution of the dye around the surface of the tank, but the dye did not travel as far as in the absence of Bond.3rd pass – The redistributed dye still did not migrate very far, and Bond was still visible within the treated area. This could be removed with continued washing.

The use of Nortron significantly reduced drying times after application, and the spray solution was more evenly distributed across the test surface. The redistribution of dye across the sprayer surface was an important observation with regards to cleaning sprayers; it is possible that residues on the surface could be removed from flat surfaces easily, but these residues may then migrate to other areas during washing where they could collect in irregular shaped areas, such as screw threads. These findings should be considered when providing guidance on sprayer cleaning.

A further observation of note was that the residues remaining on the surface of the tank during washing were largely off-white. It was postulated that a significant amount of the dye was being removed from the surface whilst the Bond remained behind. This would clearly limit the usefulness of this solution as a tracer for quantitative work to support the observations thus far.

Fluorescein + 10% Bond

To investigate whether the separation of disulphine blue and Bond during washing was an artefact of the dye used, a study was conducted using fluorescein in place of disulphine blue. From observation, fluorescein was wetted up and transported more easily than disulphine blue, and there was clearly no improvement in the technique. This tracer was therefore not considered further.

Adjuvants + solvents

The findings from using Nortron and Bond were encouraging in that the spray solution could be applied evenly, and it dried relatively rapidly. However, as Nortron was xylene based, and therefore potentially harmful, an alternative solvent would be preferable. The suitability of an ethanol/methanol mix as a diluent was therefore investigated. In addition, the suitability of other adjuvants in combination with the organic solvent as a diluent was investigated. Although other adjuvants had been investigated previously it was postulated that any effect of the adjuvants on retaining the dye on the tank surface may be augmented using the larger surfaces enabling significant differences to be observed. The adjuvants and the effect of adding an ethanol/methanol mix are summarised below:

Activator 90 – Loses wetting properties when sprayed, so the droplets gather and ‘run’Bandrift – Incompatible; coagulates

SID 5 (2/05) Page 9 of 25

Bond – Incompatible; de-natured the latexDesigner – Resulting mix a milky colour; would be poor for analysisPhase II – OKSprayfast – Incompatible; de-naturedTorpedo II – OK, but loses some wetting properties

On the basis of the above findings, only Phase II and Torpedo II were investigated further. It became apparent that Phase II adversely affected the performance of the applicator resulting in a patchy distribution of the solution, but Torpedo II could be applied evenly. For both the adjuvants, some residues of the mixture remained on the surface after an initial wash. Continuous washing with water could not remove these residues if the distance between the lance and surface was greater than c. 20 cm. This indicates that there may be a critical pressure below which an increase in water volume is of no consequence.

The overall conclusions of the above experiments was that the volumes of water applied were so large in relation to the mass of dye that the highly soluble tracer compounds were easily, and preferentially removed from the sprayer surface leaving any ‘sticker’ on the surface; the problem lay with the solubility of the tracers and not the characteristics of the sticker. The use of Bond allowed some visual observations to be made:1. Compounds on sprayer surfaces may be redistributed rather than removed during cleaning depending on the

water volume used/duration of cleaning.2. There was a reduction in efficiency of cleaning with increasing distance between the lance and target.3. There was a reduction in efficiency of cleaning with digression from a perpendicular impact of the water.

These qualitative observations were used to prioritise further work areas, but to provide conclusive results, quantitative data (i.e. a tracer) were still required.

The possibility of using a water-soluble paint as a tracer was therefore considered. The material selected was a commercial formulation, Coolglass™, pbi Home & Garden Limited 2005, designed to be painted on the external surface of greenhouse glass to limit the transmission of solar radiation during periods of intense light and heat. The finely milled powder was prepared as a suspension as per the manufacturers directions at the rate of 67g L -1

water. Coolglass concentrations were determined using a Cecil CE1010 spectrophotometer at a wavelength of 680nm. The calibration curve was linear (r2 = 0.99) within the range of 15.6 mg L-1 to 500 mg L-1 using six standards.

Coolglass™ was identified as a suitable tracer with regards to sprayer cleaning investigations and this compound was used for all subsequent quantitative investigations.

2.2 Investigation into factors affecting cleaning efficiencyThe distance and angle to target were investigated further to assess the extent to which they limit the cleaning efficiency.

The Coolglass™ solution was vigorously agitated before use to ensure a homogeneous mix. The solution was applied to the tank surface using a 50 mm paint-brush; the quantity applied was determined gravimetrically by weighing the brush and container before and after application. The tank was air dried for 2-3 hours before being washed with a cold water pressure washer for 30 seconds at distances of 0.25, 0.5, 1 and 1.5 m between the end of the lance and the treated surface. In addition, for each of these distances, the lance was held to create an impact angle of 90° and 45°. Four replicates per test were used. All the washings were collected and the volume was quantified gravimetrically. A sub-sample was analysed for Coolglass concentration and the mass of Coolglass removed was calculated by multiplying the concentration by volume.

SID 5 (2/05) Page 10 of 25

Figure 2 The effect of distance to target and angle of water jet on cleaning efficiency

The results clearly demonstrate the reduction in efficiency of cleaning with increasing distance between the lance and the target. Deviation from a perpendicular impact of the jet also reduced the cleaning efficiency and this was more marked with a distance to target of 1 m or greater where there was a relative loss of efficiency of over 50%.

The lower cleaning efficiency at the greater distance is probably a factor of the lower pressure as the water impacts the surface. It was therefore postulated that a low pressure washer may have significantly less cleaning power than a high pressure washer. Investigation of this theory was of particular relevance to the aim of this study, as in-field cleaning kits are powered by low pressure pumps, typically in the region of 10 bar; this compares to an industrial pressure washer in the farmyard of c. 100 –150 bar. The efficiency of different cleaning methods was therefore investigated. With the exception of the brush, the distance from lance to target was c. 0.4 m. The Coolglass was applied, collected and analysed as described above.

Table 2 Cleaning systems investigated

Cleaning system Equipment used Typical water flow rate (L min-1)

Hot high pressure jet (Hot pw) KEW 16A2V 8.10Cold high pressure jet (Cold pw) KEW 16A2V 8.75

Cold low pressure jet (In-field) Hardi 843197 pistol type 60L 5.75Cold hand brush (Brush) Hozelock Car care brush 10.2

The results illustrate that the in-field cleaning kit was inefficient compared to the other cleaning systems tested and this effect was more marked when the tracer was left to dry for more than 1 day. The hot pressure washer may provide a better clean than the cold pressure washer when the residues have dried for longer (Figure 3).

Figure 3 Efficiency of the different cleaning systems

SID 5 (2/05) Page 11 of 25

2.3 Field scale experimentIt was intended that the results of the small and medium scale experiments would provide a tracer suitable for use at the field scale where further testing could be conducted. However, due to the initial difficulties in identifying a suitable tracer, and the fact that the final ‘tracer’ was an emulsion, the field scale experiments were not as extensive as intended.

To pilot the collection method to be used in the field, a trial was conducted using disulphine blue and 10% Bond as the spray solution. The spray solution was chosen on the grounds that it was the best available at the time. A Hardi 300L mounted sprayer was driven in a circle, both clockwise and anti-clockwise with a boom height of 75 cm over short grassland. The sprayer was cleaned with a cold pressure washer and the washings collected on a tarpaulin. It was noted that dye was still dripping from the boom after it was cleaned for a ‘typical’ length of time; the boom was therefore cleaned a second time and these washings collected separately. The fact that dye was still visibly dripping off the boom after the first clean may be an artefact of the redistribution of surface residues, as was identified in the smaller scale experiments.

The contribution of each discrete area to the total loading is presented in Figure 4. In this test, the boom contributed to over 50% of the total loading with residues from the tank, wheels, and tractor being approximately equal. The tarpaulins proved to be a practical method for collecting washings from the sprayer, and they would be suitable for use in the field.

Figure 4 Contribution of each discrete area to total loadings in the washings

3 FIELD MONITORINGTo provide definitive data on the potential environmental impact of cleaning the external surfaces of sprayers, quantitative data on the pesticide loading contained in the washings was required. It is known that the boom and the back of the sprayer will be more contaminated than the rest of the sprayer (Ramwell et al., 2004; Arlemo, 2002). It therefore follows that pesticide loadings in washings from discrete areas of the sprayer will differ, as would any associated environmental risk. Quantifying the loadings from discrete areas would provide more accurate data to develop mitigating strategies, for example, whether parts of the sprayer should be cleaned in the field before returning to the farmyard.

3.1 Field methodsThe sprayer was sectioned into four areas: the boom and the back of the sprayer, the rear wheels, the spray tank, and the cab/rest of the sprayer. External residues on each of these four areas were quantified by wiping three 10 x 10 cm areas using a cotton swab wetted with methanol (5 ml) to form a composite sample; it was intended that this should account for some of the variation in external residue deposits that could be expected to occur. Two tarpaulins were laid on the ground under either side of the boom. The boom was lowered and left half folded for practical purposes. The spray operator cleaned the boom following normal farm practices. The time taken to clean the boom was noted, and the flow rate of the washing device was measured. A sub-sample (1L) of the washings was taken. This process was repeated for the rear wheels, spray tank, and cab, using a different tarpaulin for each of the discrete areas. The sprayer was driven onto and off the tarpaulins as required. After

SID 5 (2/05) Page 12 of 25

washing, remaining external residues were quantified using cotton swabs as described above. Samples were stored at 4°C prior to analysis.

3.2 Chemical analysis3.2.1 WaterSurface water (100 ml) was partitioned with 2 x 20 ml of dichloromethane, draining the solvent layer into a 50 ml volumetric flask and making up to volume with additional dichloromethane. A portion of the extract was concentrated to dryness and redissolved in ethyl acetate for analysis of azoxystrobin, chlorothalonil, flusilazole, pendimethalin, pirimicarb and tebuconazole by GC-MS. For carbendazim and isoproturon, a portion of the extract was concentrated to dryness and redissolved in methanol; this was analysed by LC-MS-MS. Each batch contained a determination of at least one control sample extract (HPLC grade water) and one fortified control extract that was prepared concurrently with the samples. The fortified control extract was prepared by adding a known amount of a mixed fortification solution containing all compounds of interest. The method was validated for each analyte by fortifying four control samples at 0.5 µg L-1.

All pesticide standards were supplied as neat materials with certified purities ranging from 98 to 99.7% . A mixed stock solution at 20 µg ml-1, containing azoxystrobin, chlorothalonil, carbendazim, flusilazole, isoproturon, pendimethalin, pirimicarb and tebuconazole was prepared in ethyl acetate and methanol and eight calibration standards were produced in the range of 0.001 µg ml -1 – 1.0 µg ml-1. Linear regression was used to determine the best-fit straight line through the origin for the plot of peak area versus concentration for azoxystrobin, flusilazole, pendimethalin, pirimicarb, carbendazim, isoproturon and tebuconazole (correlation coefficients ranging from 0.98 to 1.00), and polynomial regression for chlorothalonil. The mean recovery and RSD values are tabulated below (Table 3).

Table 3 Recovery levels for water samples

Azo

xyst

robi

n

Chl

orot

halo

nil

Flu

sila

zole

Pen

dim

etha

lin

Piri

mic

arb

Teb

ucon

azol

e

Isop

rotu

ron

Car

bend

azim

Fortification level (µg/L) 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5Mean recovery (%) 84.0 80.7 76.8 74.5 76.9 82.1 88.4 65.4

% RSD 8.9 7.2 6.0 3.6 5.2 6.1 4.4 4.4

Originally, all concentrations above the lowest calibration solution, equivalent to 0.1 µg/L for each analyte were reported. However, the concentrations detected were higher than anticipated, so the calibration range was updated to reflect these higher concentrations. In the updated method, all concentrations of pirmicarb, pendimethalin, flusilazole, isoproturon and carbendazim above 0.5 µg/L were reported and all concentrations of chlorothalonil, tebuconazole and azoxystrobin above 2.5 µg/L were reported.

3.2.2 Swabs Swab samples, which consisted of three cotton pads, were shaken vigorously with ethyl acetate (100 ml) for 5 minutes followed by sonicating for 30 minutes. An aliquot of the extract was analysed by GC-MS for azoxystrobin, chlorothalonil, flusilazole, pendimethalin, pirimicarb and tebuconazole and by LC-MS-MS for carbendazim and isoproturon. Each batch contained a determination of a least one control sample extract (three methanol-washed cotton pads) and one fortified control extract that was prepared concurrently with the samples. The fortified control extract was prepared by adding a known amount of a mixed fortification solution containing all compounds of interest. The method was validated for each analyte by fortifying four control samples at 0.1 µg ml-1.

A mixed stock solution at 20 µg ml-1, containing azoxystrobin, chlorothalonil, carbendazim, flusilazole, isoproturon, pendimethalin, pirimicarb and tebuconazole was prepared in ethyl acetate and methanol and six calibration standards were produced in the range of 0.02 µg ml-1 – 1.0 µg ml-1. Linear regression was used to determine the best-fit straight line through the origin for the plot of peak area versus concentration for azoxystrobin, flusilazole, pendimethalin, pirimicarb, carbendazim, isoproturon and tebuconazole (correlation coefficients ranging from 0.98 to 1.00), and polynomial regression for chlorothalonil. The mean recovery and RSD values are tabulated below (Table 4). All concentrations above the lowest calibration solution, equivalent to 0.02 µg ml -1 for each analyte were reported.

SID 5 (2/05) Page 13 of 25

Table 4 Recovery levels for swab samples

Azo

xyst

robi

n

Chl

orot

halo

nil

Flu

sila

zole

Pen

dim

etha

lin

Piri

mic

arb

Teb

ucon

azol

e

Isop

rotu

ron

Car

bend

azim

Fortification level (µg/ml) 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

Mean recovery (%) 86.3 72.4 72.5 82.5 78.3 69.0 125.6 100.4

% RSD 2.9 3.5 2.5 1.9 1.8 1.9 3.0 1.5

3.2.3 Mud A 20g portion of mud was extracted with acetone/water and filtered through a GF/A filter paper in to a 250 ml separating funnel containing 12g of anhydrous sodium sulphate. The filtered extract was partitioned with 60 ml of dichloromethane and concentrated to give a final volume of 2 ml. The concentrated extracts were analysed by GC-MS for azoxystrobin, chlorothalonil, flusilazole, pendimethalin, pirimicarb and tebuconazole and by LC-MS-MS for carbendazim and isoproturon. Each batch contained a determination of a least one control sample extract (LUF-Speyer 2.2 soil) and one fortified control extract that was prepared concurrently with the samples. The fortified control extract was prepared by adding a known amount of a mixed fortification solution containing all compounds of interest. The method was validated for each analyte by fortifying 4 control samples at 0.01 mg kg-1.

The calibration curve was established as described in Section 3.2.2. The mean recovery and RSD values are tabulated below (Table 5).

Table 5 Recovery levels for mud samples

Azo

xyst

robi

n

Chl

orot

halo

nil

Flu

sila

zole

Pen

dim

etha

lin

Piri

mic

arb

Teb

ucon

azol

e

Isop

rotu

ron

Car

bend

azim

Fortification level (mg/kg) 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Mean recovery (%) 124.5 65.7 83.6 95.8 77.4 92.5 67.8 64.3

% RSD 2.2 7.2 3.2 6.5 4.4 4.8 3.7 3.6

3.2.4 LC-MS-MS Instrumental conditionsThe LC-MS-MS instrumental conditions used for water, swab and mud analysis are shown below:

Instrument Sciex API 2000 (PE/ Applied Biosystems)Inlet [HPLC System] Binary Pump 1100 Series (HP/Agilent)Autosampler 1100 Series (HP/Agilent) Vacuum Degasser 1100 Series (HP/Agilent)System Controller 1100 Series (HP/Agilent)Column Waters Atlantis dC18 150 x 2.1 mmColumn temperature AmbientMobile Phase A 10 mM Ammonium AcetateMobile Phase B MethanolGradient Time % A % B

05

151625

9010109090

1090901010

Flow Rate 0.2 ml/minIon polarity Positive

SID 5 (2/05) Page 14 of 25

Injection 5 - 50 µl (or higher)Detection mode MS-MSMRM transition 1 192.1/160 for carbendazim; 192.1/132 for carbendazimMRM transition 2 207.1/165 for isoproturon; 207.1/72 for isoproturon

3.2.5 GC-MS instrumental conditionsThe GC-MS instrumental conditions used for water, swab and mud analysis are listed below:

Autosampler Agilent 7683B SeriesGC Agilent 6890N with electronic pressure controlDetector HP 5973 ‘Inert’ MSDInjector Split/splitless operated in splitless mode30 m (nominal) x 0.25 mm i.d., 0.25 µm film thicknessInjector Temperature 250°CDetector Temperature 300°COven temperature Program Initial temperature 100°C held for 1 min, then 20°C/min until

300°C, held for 5 minsElectronic pressure control to provide a constant flow of helium at approximately 0.7 mL/minInjection Volume 2 µlIonisation mode Electron Impact (EI) positive modeMS ions monitored m/z = 344 for azxoystrobin m/z = 266 for chlorothalonil

m/z = 233 for flusilazole m/z = 252 for pendimethalinm/z = 166 for pirimicarb m/z = 250 for tebuconazole

3.3 Data analysisThe mass of pesticide removed from washing each discrete area was calculated from the product of concentration and water volume used. The volume of water used was calculated from the product of flow rate and time spent cleaning. Where a bucket and brush was used, the amount of water used was calculated from that in the full bucket minus the volume remaining in the bucket after cleaning.

A predicted environmental concentration (PEC) was calculated by dividing the pesticide mass from each discrete area by the volume of water in a FOCUS-defined stream (100 m long, 0.3 m deep and 1 m wide = 30,000 L). This was compared to the maximum predicted environmental concentration calculated using the FOCUS surface water model (Step 1) assuming application in northern Europe and using a maximum application rate; the application rates, scenarios used, and the resulting PECs are shown in Table 6.

Table 6 FOCUS Step-1 maximum PECs in surface water

Active ingredient Application rate (g ha-1) Crop type PEC (µg L-1)

Azoxystrobin 250 Cereal 1.30Carbendazim 500 Oil seed rape 0.15Chlorothalonil 1500 Oil seed rape 0.20Flusilazole 200 Cereal 1.00Isoproturon 2500 Cereal 1.30Pendimethalin 2000 Potatoes 0.55Pirimicarb 280 Oil seed rape 1.70Tebuconazole 250 Cereal 1.20

Linear regression was used to investigate the relationship between the mass of compound in the washings and influential factors. Analysis of variance was used to assess differences where appropriate.

3.4 Results3.4.1 Contribution of each discrete area to total pesticide loadingWhen considering the mass of pesticide each discrete area contributed to the total loading, the boom and spray tank accounted for over 80% of the total load, and there was no significant difference in the pesticide loading from the cab area or the rear wheels of the sprayer (Figure 5). The different active ingredients did not significantly influence contribution to the total loading.

SID 5 (2/05) Page 15 of 25

Figure 5 Contribution of each discrete area to the total pesticide loading showing the mean ± 1 se

3.4.2 Measured pesticide concentrationsThe concentration of the pesticides in the washings varied between compounds and between the discrete areas. The range in concentrations for the four discrete areas is illustrated in Figure 6 and Figure 7 (note the different scale!). In general, pesticide concentrations in the washings from the boom and tank were approximately one order of magnitude greater than concentrations from the cab or wheels. Three pesticides (isoproturon, tebuconazole and chlorothalonil) were detected at consistently higher concentrations than the other compounds investigated regardless of the area sampled; pirimicarb had the lowest concentrations.

Figure 6 Pesticide concentrations in the washings from the boom and tank showing the median (symbol), the 25th and 75th percentiles (box), and the non-outlier min & max (whiskers).

SID 5 (2/05) Page 16 of 25

Figure 7 Pesticide concentrations in the washings from the cab and wheels showing the median (symbol), the 25th and 75th percentiles (box), and the non-outlier min & max (whiskers).

Mud samples from the rear wheels were also collected when there was a liberal coating, which occurred on three occasions. The quantity of pesticide in the mud ranged from below the limit of calibration to 2.8 mg kg -1. To put these values into context, the mass of pesticide contained in the mud was compared to that from the wheel and the sprayer as a whole. Assuming a total quantity of 5 kg of mud, the mass of pesticide contained in the mud represented a median value of 3.5%. The average value was 20% but this value was skewed by two samples that contained relatively high levels of pendimethalin; excluding these two samples gave an overall mean of 6.5% for the level of pesticide in the mud compared to that from the sprayer as a whole. When comparing the pesticide mass in the mud to that contained in the water phase of the washings from the wheel, the mud frequently contained a higher mass of pesticide and this could be one or two orders of magnitude greater.

3.4.3 Predicted environmental concentrationsThe mass of pesticide removed from each discrete area was summed to give a total mass per sprayer for each compound investigated. This mass was divided by the volume of water in a FOCUS-defined ditch (30,000 L) to give a predicted environmental concentration, assuming that all the pesticides removed from the sprayer were transported to the ditch with no dissipation. The PECs from this calculation and from the FOCUS surface water model (Step 1) are illustrated below for each compound (Figure 8). The PECs from washing sprayers were all lower than the PECs from the surface water model, and with the exception of pendimethalin and tebuconazole, the PECs from washing sprayers were approximately two orders of magnitude less.

Sixty percent of the samples contained pesticide residues in excess of 0.1 µg L-1 with 48% of these emanating from the boom and tank, but this was compound dependent. The majority (80%) of pirimicarb, flusilazole, azxoystrobin and carbendazim concentrations were below 0.1 µg L -1, and the higher concentrations could be attributed to washing the boom alone with the exception of azxoystrobin. For isoproturon and chlorothalonil, washings from the cab could exceed the 0.1 µg L-1, but to a lesser extent than washings from the boom or tank.

SID 5 (2/05) Page 17 of 25

Figure 8 Predicted environmental concentration in a FOCUS-defined ditch of the pesticides arising from sprayer cleaning (crosses). The diamonds are PECs from the FOCUS surface water model (Step 1).

One mitigating strategy to reduce the environmental impact of removing residues from the external surfaces of the sprayer would be to clean the sprayer in the field. Any compound used in the field will have already been risk assessed during the approvals process, and this assessment will assume a range of given application rates. An indication of the relative magnitude of the residues removed from the sprayer could be gained by calculating the area of land that the pesticide mass could treat and comparing this to a typical application rate for that product (chlorothalonil: 1.0; pirimicarb: 0.13; pendimethalin: 2.0; flusilazole: 0.2; tebuconazole: 0.25; azxoystrobin: 0.25; isoproturon: 2.0; carbendazim: 0.25 kg ha-1). A comparison of the area that the residues could treat with the area on which the washings would fall provides some indication of whether the residues were at levels that could adversely impact on the environment.

Over 90% of the samples contained residues that were sufficient to treat an area of less than 1 m2 and the largest area that could be treated was c. 4 m2 (Figure 9), which was a sample from the boom; tebuconazole and isoproturon could treat the largest areas. Even if a boom remained half-folded during washing, the area onto which the washings would fall would commonly be in excess of 10 m2, and that is excluding the area beneath the rest of the sprayer. It is therefore unlikely that cleaning the sprayer in the field will have any adverse environmental impact, as long as suitable precautions are taken to prevent the washings water flowing into surface waters or similar.

SID 5 (2/05) Page 18 of 25

Figure 9 The land area that could be treated with residues contained in the washings

3.4.4 Measured residues on the external surfaceResidues on the external surfaces of the sprayer were sampled using methanol-wetted swabs prior to and after washing in order to provide an estimate of the total residues available for removal and the efficiency of the cleaning regime. The quantity of pesticide remaining on the sprayer after washing was compared to the initial mass, and the results are presented below (Figure 10). There was a lot of scatter in the data, and no general observations could be identified. Samples taken from a single site on a single sprayer could contain variation between compounds ranging from 7% to 200%, but there was no consistency. For some samples there was an apparent increase in mass of pesticide on the sprayer after washing. This is likely to be a reflection of the natural variability that occurs when spray is deposited on the external surfaces of the sprayer, particularly where residue levels are low, and any small change in the dose on the surface could result in a large percentage change.

Figure 10 Residues remaining on the sprayer after washing as a percent of initial mass

SID 5 (2/05) Page 19 of 25

x Boom Tank Wheel Cab

31500

3.4.5 Pressure washer waterPesticides were detected at all farms in the water collected from the pressure washer to measure flow rate, except where only the brush and bucket was used. The presence of a compound appeared to be compound dependent, and whilst there were few positive detections for chlorothalonil, pirimicarb, and azxoystrobin, there were a number of detections above 0.1 µg L -1 for isoproturon and flusilazole (Figure 11). The high values of pirimicarb and flusilazole were from a single farm on the same sampling event, but it is difficult to explain the finding given that there was no record of use of flusilazole in that season, and other compounds that had been used were not detected in the pressure washer water. On the farm where the bucket and brush was used, seven out of the eight compounds investigated had a record of use, but there was no detection of the compounds in the water collected in the bucket for the control sample. There was no consistency in the findings for other farms using pressure washers in that sometimes residues were detected where there was no record of use in the past season, and, conversely, some pesticides that had a record of use were not detected. However, for farms using a pressure washer, where there was a record of use of isoproturon, there were relatively high concentrations of isoproturon in the pressure washer samples (range 1.18 – 3.97 µg L-1).

Figure 11 Pesticide concentrations in pressure washer water

4 DISCUSSIONThe field monitoring studies clearly demonstrated that the boom and back of the sprayer were the most contaminated areas of the sprayer, contributing approximately 80% of the total pesticide loading. These findings are in accord with a previous study quantifying external residues on sprayers (Ramwell et al., 2004) that investigated the presence of thirteen compounds on thirteen different sprayers. However, the results were not wholly consistent with the findings of the Cherwell study (Mason et al., 1999) which identified the rear wheels as a significant contributor to the total pesticide loading during washing, as well as the back of the sprayer and the boom. The results of the current study would indicate that, on the whole, the rear wheels do not contribute significantly to the total pesticide loading compared to the boom/back of sprayer as demonstrated by the consistently lower contribution for a number of different sprayer types, irrespective of compound.

The quantity of pesticides contained in mud on the rear wheels was also lower in the current study (2.2 mg kg-1 of isoproturon) than in the Cherwell study (8 to 10 mg kg -1 isoproturon). It is possible that field conditions and/or soil type may have contributed to these differences. Although the current study investigated a larger number of sprayers and compounds, there were only three occasions when the field conditions were muddy, and the longer time lag between spraying and sampling may have allowed more sorption of the pesticides to the mud. The contribution of mud transported to the farmyard to the total pesticide loading may be more variable than the results of the current study were able to demonstrate. Indeed, other workers found that the presence of mud around the wheel area can result in massive variations in residues removed during washing (Holst et al., 2001). The contribution of mud to the total pesticide loading will tend to be seasonal, and this may be mirrored by different pesticide usage. For example pendimethalin and isoproturon are autumn herbicides and they are therefore applied when field conditions will have a higher tendency to be muddy. The results of the Cherwell study may therefore not be representative of general spraying conditions for a variety of compounds, and this is substantiated by the results of the current study, but the Cherwell results may provide an example of the extent to which contamination can occur under a more “worst case” scenario, and/or where the soil type is more clayey.

SID 5 (2/05) Page 20 of 25

Evaluating the PECs from the FOCUS surface water model and the PECs calculated in this study indicated that residues from washing sprayers are comparatively low. If the drinking water standard of 0.1 µg L-1 is used as an endpoint then 60% of the samples contained pesticide residues in excess of 0.1 µg L-1 with 48% of these emanating from the boom and tank. However, this assumes that all the residues contained in the washings enter a surface water body with no dissipation during the transport process, or in the receiving water body.

In an attempt to explain the behaviour of the compounds observed in the current study, the pesticide mass in the washings, and the pesticide mass as a percent of the total residues available, were compared to the solubility of the compound, the application rate, and the water : octonal partition coefficient. However there were no significant relationships. This may partly be a consequence of the large number of factors that can affect the deposition of residues on the external surface in the first place, such as formulation, wind speed, ground speed, crop type and height, which would negate the identification of any relationship with the physico-chemical properties of the compounds. Consequently, it was not possible to predict accurately the behaviour of compounds outside those investigated in the current study. Nevertheless, as the current study considered compounds with a range in properties for a number of sprayers and sprayer types, the results can be considered representative of reality. The consistency of the findings in relation to the contribution of the boom/back of sprayer and the spray tank to the total pesticide loading would indicate that these findings are likely to be representative for all compounds, including those not investigated here.

As a precautionary measure it may be advisable to wash the sprayer down in the field, away from water courses. The study has demonstrated that the residues represent only a very small fraction of typical application rates, thus overdosing will not occur if managed appropriately. The areas that could be treated by the pesticide mass contained in the washings were similar to a study by Ganzelmeier (1998) for the compounds common to both studies (isoproturon, pirimicarb and tebuconazole), although for some compounds Ganzelmeier (1998) detected sufficient residues to treat areas of over 10 m2, but he concluded that overdosing would not be expected. Whilst cleaning sprayers in the field is likely to negate any environmental concerns compared to cleaning the sprayer in the farmyard, this is assuming that residues are effectively removed in the field. The experimental work with tracers indicated that onboard low-pressure washers removed approximately one third of the available residues, assuming that pesticides are removed to the same extent as Coolglass. The remaining residues could therefore be transported back to the farmyard and these may then be available to be removed during rainfall events. However, residues could be removed more effectively in the field with a brush, thus minimising contamination on return to the farmyard.

The experimental studies clearly demonstrated the comparable efficiency of the brush and the high pressure washer using Coolglass as a tracer. In the current study, two farms used a brush and bucket method, with one following with a hose rinse. The mean mass of pesticide detected in the washings from these farms included the lowest recorded (0.4 mg) and 7.8 mg; this compares to an average mass of 18 mg. The results from a previous study quantifying external residues on working sprayers under field conditions also supported the findings of the brush efficiency (Ramwell et al., 2005b). The study considered the percentage of samples below the limit of detection (LOD) on each sprayer in relation to the cleaning method used. The findings are presented in Table 7. Although there were insufficient samples for robust statistical analysis, the results indicated that the onboard hose for infield cleaning was the least efficient cleaning method, and that the brush had a comparable performance to the pressure washers.

Table 7 Percentage of samples <LOD for individual sprayers

Samples <LOD (% of total) Cleaning method Farm

67 Steam pressure a53 Hose b57 Hose b38 Onboard hose c70 Hot pressure d65 Onboard hose + brush e69 Onboard hose + brush e

Source: Ramwell et al., 2005b

If sprayers are cleaned in the field, the available water supply is likely to be limited by the size of the clean water tank. One disadvantage of the brush was that the flow rate was almost twice that of the onboard lance without a brush; thus there is half the time available to clean the sprayer before the water supply is consumed. It may therefore be advantageous for manufacturers to consider reducing the flow rate to the brush to some extent.

SID 5 (2/05) Page 21 of 25

If, due to time or water constraints, it is not possible to clean the entire sprayer in the field, it may be advisable to ensure that, as a minimum, the boom/back of the sprayer and the spray tank are cleaned thoroughly in addition to removing mud. However, there are potential health impacts of leaving residues on the sprayer (Ramwell et al., 2005a) thus it would be advantageous to ensure that the entire sprayer is cleaned. Furthermore, if sprayers are cleaned regularly, this may prevent the build up of residues, thus minimising any health and/or environmental impacts associated with the residues.

The reduced cleaning efficiency of water jets as the distance of nozzle to impact surface increases could have implications for the cleanliness of different parts of the sprayer. The boom on a sprayer is commonly stored folded relatively high up during transport. To effectively clean the boom, it would have to be lowered and unfolded. This would also increase accessibility to the different surfaces of the boom for cleaning, further enhancing the effectiveness of the clean. Moreover, the need to open and unfold the boom to ensure an efficient clean would have implications for washing down sprayers over restricted catchment areas such as those used for drive-over biobeds, and, for an offset biobed, it may be necessary to drive the sprayer on and off the bunded area to clean, for example, half the boom at a time, unless a very large bunded area is created, or one specifically for the boom. It may therefore be more practical to clean the sprayer, or at least the boom, in the last field of use. The experimental data also illustrated the possibility of re-distribution of the residues around the machine during cleaning. Over recent years, manufacturers have taken sprayer cleaning into consideration in the design of the machinery, and many new sprayers have covers over the intricate pipe work to provide a smooth surface for cleaning after pesticide application. However, unless suitable care is taken, the residues from this smooth surface may just be washed off and settle in difficult-to-reach places. This may be more so with a pressure washer than a brush, and a further advantage of the brush is that it is less likely to result in the redistribution of residues across the sprayer surface. However, residues may become entrained in the brush that would then allow the transfer of residues from contaminated areas such as the boom, to less contaminated areas such as the cab. It may therefore be necessary to rinse the brush after cleaning highly contaminated areas. Alternatively, the least contaminated area could be washed first, as long as the water supply is carefully managed, and there is plenty of water left to clean the boom.

In considering whether a high pressure washer or a brush is the most appropriate cleaning method, another factor to consider is the human effort required in resisting the recoil effect of directing the water onto a hard surface, particularly given that the closer the lance the better the clean. There may actually be less effort involved in using a brush, but this could depend on the individuals involved, and the human factor must be taken into consideration; if the operator starts tiring, they may be tempted to reduce the length of time for which the sprayer is cleaned. Other human factors to consider when choosing the most appropriate cleaning method are health risks, such as electrocution if the pressure washer is not diesel-powered, and falling from heights in an attempt to clean the top of the sprayer, if no purpose-built platforms to access the upper parts of the sprayer are available. A recent study by HSE has shown that residues can be transferred to the operator during washing, but this was the case for all cleaning methods investigated (hose + brush, high pressure, and hose).

A comparison of the residues on the sprayer surface prior to and following washing, as determined by the swab samples, showed an apparent increase in residues after cleaning on some occasions, particularly where residue levels were initially low. This indicated that the sampling method may not have covered sufficient area to encompass typical variation in the residues. The results were therefore not suitable to assess the efficiency of the cleaning method quantitatively. However, a comparison of the total mass of residues on the sprayer, to the pesticide mass contained in the washings indicated that in 90% of the samples, the swabs were an overestimate estimate of the pesticide mass removed during washing, thus if swab data are used as an estimate of residues contained in washings this represent a worst case scenario.

The presence of pesticide residues in the water sampled from the pressure washer was not expected. The probability of contamination during analysis was minimised by flushing the spectrometers through with solvent after calibration, and running the pressure washer sample first. On a number of occasions, the mass of a pesticide in the pressure washer water was greater than the mass collected in the washings from the sprayer, as calculated from multiplying the concentration by the volume of water used. However, this was largely where concentrations detected were very low (< 0.1 µg L-1) and any small differences in concentrations, either with or without normal analytical deviations, would make a relatively large percentage difference, and, for the purposes of this study, all the data were used as reported. It is possible that the farmyard was contaminated with low levels of pesticide residues, and, as the analytical method was designed to detect very low concentrations, these residues were detected. It was common practice to put the lance on the ground whilst, for example, the operator assisted in moving the tarpaulins for collecting the washings, and, as the flow rate was measured after washing, residues may have been transferred to the lance. The presence of low level residues on the ground around the farm is further supported by findings during the developmental stage of the analytical methods. A Landrover that had been used for general farm visits was washed and the washings collected to provide ‘blank’ water for analytical development. Some residues were detected in the washings, primarily below 0.05 µg L-1. This indicates that very low levels of residues may be present in some farmyards and environs.

SID 5 (2/05) Page 22 of 25

The experimental phase of this study provided important data with regards to factors affecting the efficiency of different cleaning methods, and, these are factors over which the operator can have complete control. The efficiency of the brush compared to the pressure washer, and the inefficiency of the onboard lance were key findings, as were the effects of angle and distance to target for the pressure washer. With regards to the fieldwork, the data illustrated that, on the whole, the environmental impact of cleaning sprayers is relatively low and any risk may be negated by cleaning the sprayer in the field away from water courses. Drive-over biobeds of the dimension 5 x 7 m would be unsuitable to allow efficient cleaning of the boom/back of sprayer, but the washings from other areas of the sprayer could be collected by such a system, or an offset biobed if the bunded area was large enough.

The farmers participating in the study all showed an interest and understood the need to provide scientific data on the environmental impact of cleaning sprayers. The farmers were all keen to ensure that their practices did minimise any adverse effect on the environment and they would welcome advice arising from the study. It was noted by some that pressure washers onboard sprayers for in-field cleaning performed poorly compared to high pressure washers, to the extent that they were not really worthwhile. It is also worthy of note that there was interest in biobeds, and some were considering installing them. However, concern remained about what to do with the biobed substrate once it was ‘spent’.

5 CONCLUSIONS

Highly water soluble compounds are largely ineffective tracers for the investigation of external sprayer cleaning, even when mixed with ‘stickers’. This may be due to the inherent solubility of the tracers in conjunction with the relatively high volumes of water used rather than any characteristic of the stickers.

CoolglassTM provides sufficient resistance to removal during cleaning to represent a suitable tracer for investigations into external sprayer cleaning at the ‘laboratory’ scale. It has the advantage of not requiring special storage facilities etc. associated with compounds registered as pesticides, although a full risk assessment of its use is still required. However, it may not necessarily be appropriate for field-scale experiments.

Tracers that our highly visible on the test surface allow observations during the cleaning process to made which can provide useful information.

Residues may be redistributed around a sprayer surface rather than washed off if suitable care to avoid this is not taken.

The distance from lance to target and the angle at which the water impacts the surface significantly affects cleaning efficiency; there is a decrease in cleaning efficiency with increasing distance and deviation from perpendicular. These are parameters over which the operator can have full control.

High pressure jet washers and brushes were equally as efficient at cleaning.

Low pressure cleaning kits, as provided on sprayers for in-field cleaning, are relatively inefficient, and a brush would provide a more efficient method of cleaning.

It may be advantageous for manufacturers to consider reducing the flow rate of onboard cleaning kits with brushes to optimise the use of the water supply.

The choice of cleaning method should consider the human effort involved, and avoid one that is unnecessarily tiring as this may reduce overall cleaning times.

The boom and the back of the sprayer contributed to 80% of the total pesticide loading, thus cleaning these areas of the sprayer in the field, as a minimum, would assist in reducing contamination in the farmyard.

It is probable that the boom will need to be lowered and unfolded for it to be cleaned effectively, due to the loss in cleaning efficiency with increasing distance to target, and the angle of impact. Drive-over biobeds would therefore not be suitable for collecting washings from the boom under these circumstances, and offset biobeds would require a large bunded area.

SID 5 (2/05) Page 23 of 25

Mud transported back to the farm may also be a significant contributor to pesticides in the farmyard; this should therefore be removed in the field where possible. The contribution of mud to the total pesticide loading from farms may depend on season and soil type.

It is unlikely that cleaning the sprayer in the field will have an adverse impact on the environment.

Both the potential health and environmental risks associated with sprayer cleaning should be considered when providing advice to operators. Results from the current study and a previous study for HSE will be combined to provide guidance to operators on external cleaning.

6 Knowledge Transfer

It is now acknowledged that cleaning the external surfaces of sprayers is an important part of the spraying process, for both health and environmental reasons. However, the process of washing down a sprayer could in itself contribute to operator exposure to pesticides; consequently, the Health and Safety Executive commissioned a study to investigate the potential exposure.

It is intended that the results of the current study will be amalgamated with the findings of the HSE study to provide advice that will enable the spray operator to encompass both the environmental and health implications of sprayer cleaning.

The findings will be disseminated to sprayer operators, sprayer manufacturers and the chemical industry via channels such as the Crop Protection Agency, the Agricultural Engineers Association, the National Sprayer Testing Scheme, and trade journals such as Farmer’s Weekly. It is also intended to generate a general press release due to the current public interest in this subject area.

The research will be presented at the International Advances in Pesticide Application Conference, Cambridge in January 2006 and included in the proceedings (Aspects of Applied Biology, 77). A journal paper is being prepared for submission to Pest Management Science.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.

SID 5 (2/05) Page 24 of 25

Arlemo T. (2000). Ackumlering av avsatt sprutvatska pa traktorn och pa lantbrukssprutans bom och tank (Accumulation of spray deposits on tractor and field crop sprayer boom and tank). (In Swedish) SLU Report 241, 23 pp.

Barber JAS and Parkin CS (2003). Fluorescent tracer technique for measuring the quantity of pesticide deposited to soil following spray applications. Crop protection 22: 15-21.

Higginbotham S, Jones RL, Gatzweiler E and Mason PJ (1999). Point source contamination: quantification and practical solutions. Proc. Brighton Crop Prot Conf – Weeds, BCPC, Farnham, pp 681-686.

Fogg P (1999). Pesticide residues from spray equipment. Unpublished SSLRC Report, Cranfield University.

Fischer VP, Hartmann H, Bach M, Burhenne J, Frede H G and Spiteller M (1998). Gewässerbelastung durch Pflanzenschutzmittel in drei Einzugsgebieten (Pesticide pollution in three watersheds). (In German) Gesunde Pflanzen 50:142-147.

Ganzelmeier H (1998). Proper cleaning of sprayers, in Managing pesticide waste and packaging, BCPC symposium proceedings No 70, BCPC, Farnham, surrey, UK, pp 91-98.

Holst CD, Nielsen C and Andersen PG (2001). Developments with the internal and external cleaning of sprayers in the field of use. Aspects of Applied Biology 66: 395-400.

Ramwell CT, Johnson PD, Boxall ABA, Rimmer DA (2002). Exposure of pesticide residues on agricultural spraying equipment. Cranfield Centre for EcoChemistry Research Report No. JA6116v for the Health & Safety Executive, 58 pp.

Ramwell CT, Johnson PD, Boxall ABA, Rimmer DA (2004). Pesticide residues on the external surfaces of field-crop sprayers: environmental impact. Pest Management Science 60: 795-802.

Ramwell CT, Johnson PD, Boxall ABA, Rimmer DA (2005a). Pesticide residues on the external surfaces of field-crop sprayers: occupational exposure. Annals of Occupational Hygiene 49: 345-350.

Ramwell CT, Johnson PD, Corns, H, Boxall ABA, Rimmer DA and Sandys V (2005b). Decontamination of agricultural sprayers. Report OMS/2005/13 to HSE.

SID 5 (2/05) Page 25 of 25