Drift when applying biocides to control crawling and flying insects on walls

Abstract

Background

Insecticides are sprayed on external building walls for treatments against crawling and flying insects. These applications can lead to drift into non-target areas and thus to undesirable environmental pollution. This emission pathway needs to be considered during exposure assessments within product authorisations to assess potential environmental risks. However, now, there is only one default value for deposition that is used in all calculations based on the Emission Scenario Document of the Organization for Economic Co-operation and Development at a distance of 50 cm to the treatment area. This is not sufficient for a risk assessment.

Results

When applying a chemical barrier of 50 cm at the bottom of the building wall, wind direction had the greatest influence on drift, while changing the nozzle type had no significant effect. Compared with the measured ground sediments, the OECD default value was deemed to be realistic at a distance of 57 cm from the treatment area. When treating the entire building wall, the wind direction as well as the nozzle used show significant influence on the measured values of drift. The default value for deposition proposed for modelling environmental exposure in OECD document ESD PT18 No. 18 was exceeded. Thus, the exposure estimation might not be protective enough.

Conclusion

Drift values used for the environmental exposure assessment of biocidal products during treatments of building walls should be adapted. This is especially relevant for treatments of entire building walls, where the current default value was exceeded for all distances from the building wall. Wind direction and nozzle type can reduce environmental impact. This finding can be used as a measure to reduce unnecessary exposure in the environment in the future.

Introduction

Biocides protect humans, animals, and materials from pests, vermin, and other harmful organisms. Like plant protection products, they belong to the group of pesticides but have their own regulation [3]. Biocidal products are very diverse. There is not just a single product line or a clear-cut application area, but biocidal products are divided into 4 main groups and 22 product types [3]. This publication focuses on insecticides used to control flying and crawling insects on building walls. These insecticides belong to product type 18. This product type 18 includes insecticides, acaricides, and products against other arthropods to control cockroaches, ants, fleas, flies, mosquitoes, bugs/hornets, spiders, dust mites, bedbugs and termites [3].

Due to the diverse use of biocidal products, there is a lack of knowledge about common practices in the use of these application methods. This leads to uncertainties in quantitative assessments of emissions into environmental compartments as part of the environmental exposure assessment [21]. Consequently, quantifiable risk mitigation measures are lacking. A fundamental issue is the absence of basic drift values, such as those used in the application of plant protection products, from which respective assessments and risk measures are derived.

In the OECD (Organization for Economic Co-operation and Development) Emission Scenario Document (ESD) No. 18 in PT18 [15] two scenarios are selected to model environmental emissions for spray treatment of buildings. The first is the control of crawling insects, where a chemical barrier of 50 cm is achieved by treating the foundation of the building and a treated band of soil around the building. The second scenario reflects the control of flying insects where it is considered that the entire wall of building is treated. For both scenarios, the document specifies only a standard fraction of 10% that is released onto the ground/paved surface during outdoor spraying [15]. Information on how the drift is captured and calculated, as well as details on the application method, is not provided.

More detailed basic drift values that could be adopted from other guidelines are not available. The basic drift values issued by the Julius Kühn-Institute (JKI) for example are for the field of plant protection and refer to applications in arable, fruit, wine, field crops, etc. that cannot be compared to the biocidal treatment of building walls. However, the methodology of the data collection for these values can be used to derive specific basic drift values for the treatment of building walls. The JKI has published guideline 7–1.5 for measuring direct drift in the application of liquid plant protection products in the field [9]. The JKI guideline is identical to the ISO 22866:2005 standard [7] and to the EPA standard (EPA 1998) in many areas, such as the orientation of the treated area and the measuring area. With regard to weather conditions, the JKI guide is even stricter. The average air temperature must not exceed 25 °C during the entire trial. The average wind speed must be between 1 and 5 m s⁻¹ [9]. Therefore, the JKI guideline is more protective to derive basic drift values for environmental exposure assessment.

In this publication, the scenarios of the OECD ESD for PT 18 No. 18 and the JKI guideline for drift measurements are used as a basis to measure drift during the control of flying and crawling insects on building walls with a knapsack sprayer and different nozzles. The aim is to test the hypothesis that one default value is not sufficient to build a risk assessment for this specific application. This is the first publication to measure the drift exposure as a scientific basis for further risk assessment and to estimate the predicted environmental concentration.

Material and methods

Knapsack sprayer and Nozzle types

The REC 15 AC1 (Birchmeier Sprühtechnik AG, Stetten, Switzerland) knapsack sprayer was used to treat the entire house wall and for foundation treatment (treatment of a chemical barrier). This knapsack sprayer is a pressure-controlled battery knapsack sprayer. The pressure range is manually adjustable from 0.5 to 6.0 bar, but for these trials the pressure was set at 2.0 bar (according to nozzle approval [10]). The tank has a capacity of 15 L. In addition, the lance was extended with an extension tube according to the experiment’s setup. Two different nozzles were used to determine the variation in drift behaviour. In this study, a brass hollow cone nozzle and the nozzle ID 90–015 C were used. The brass hollow cone nozzle with a nozzle size of 1.7 mm is included as standard when purchasing backpack sprayers. Based on the flow rate of the brass nozzle, a comparable nozzle was determined. Own measurements of the flow rate with this knapsack sprayer showed that the brass nozzle had a flow rate of 0.4 L min⁻¹ at 2.0 bar. Accordingly, the nozzle IDK 90–015 C with a flow rate of 0.46 L min⁻¹ was selected using this knapsack sprayer at 2.0 bar. The nozzle IDK 90–015 C is an injector flat spray nozzle with a nozzle size of 015. This nozzle is approved for orchards and vineyards with a pressure range between 2.0 and 20.0 bar [10]. The droplet size distribution is very coarse to medium. Table 1 shows the main parameters and the droplet size spectrum according to the ISO 25358:2018 [8] at 2.0 bar (internal JKI results). In conclusion, the brass nozzle has the highest proportion of fine droplets and has a higher spray drift behaviour than the IDK 90–015 C nozzle due to these fine droplets.

Table 1 Main parameters and classification of the nozzles at 2.0 bar

Application area and procedure for measuring spray drift

The OECD ESD describes a typical private house (17.5 m long and 7.5 m wide). The wall height is 2.5 m. Nothing was reported about the type of façade [15]. As a house with these dimensions was not available for carrying out these trials, an overseas container was used as a trial area. The long side of the container was 7.55 m long and 2.45 m high and was covered with a ribbed plexiglass pane to simulate the structure of the building wall. Even though it is known that different types of façades have different properties, Plexiglas© panels were used because these panels do not absorb water and therefore represent the required worst-case scenario.

The container was used to study two scenarios for insect control by spraying: the treatment of a chemical barrier against crawling insects and a wall application against flying insects. For crawling insect control, a foundation was sprayed up to a height of 50 cm; for flying insect control, the entire wall was treated. For practical and more feasible reasons, the measuring area was limited to 1.80 m from the wall. To be able to measure spray drift up to 1.80 m distance from the container when treating a chemical barrier, trials have been carried out with two walking paths for the user. For the first path, the user walked directly in front of the container and used the "normal" lance of 60 cm length and for the second path, the user walked behind the measuring area and used the "long" lance of 160 cm length. The distance between the nozzle and the wall was always 20 cm for both paths. For the evaluation of the data, the petri dishes used as collectors of both pathways were subsequently combined. When treating the entire wall, it was not possible to use two paths, the user used only the "long" lance and walked behind the measuring area. During all trials, the user walked backwards and treated the container with even up and down movements of the lance (Fig. 1).

Fig. 1

User treated a chemical barrier (left) and the entire wall (right) with a knapsack sprayer. Petri dishes as collectors are placed in the measuring area

To map a treatment against crawling and flying insects, the average application rate was 100 mL m⁻². This application rate was controlled by the running time of the user. Considering the size of the application area, the application rate and the flow rates of the nozzles, the user needed 49 s to treat the chemical barrier and 240 s to treat the entire wall with the IDK 90–015 C nozzle and 57 s to treat a chemical barrier and 277 s to treat an entire wall with the brass nozzle. Petri dishes used as collectors had a diameter of 145 mm and were placed close together in rows. Ten collectors per row were used and placed in 8 rows. This resulted in a measured distance of 15, 29, 43, 57, 71, 85, 99, 113, 127, 141, 155, 169 and 183 cm, depending on the pathway. The spray liquid used was water with Pyranine (CAS number 6358-69-6) as the tracer dye at a concentration of 5 g L⁻¹. Pyranine is a green-yellow, powdery sodium salt (trade name: Pyranine 120%, colour index: Solvent Green 7) and has a recovery rate of almost 100% [6]. Tank samples were taken during the trials to check the application rate and to determine whether the tracer concentration was stable throughout the application. To measure the influence of the wind direction, the trials were carried out in three wind directions: parallel wind direction (WSW), orthogonal wind direction (SSE) and wind shadow of the container (NNW) (Fig. 2). Five minutes after spraying, the collectors were closed and immediately protected from light. The analysis of the tracer took place in the laboratory with a fluorometer (RF-6000, Shimadzu Duisburg, Germany). In addition, collectors were set up outside the measuring area to determine the blank value. According to JKI Guideline 7–1.5 [11] and ISO 22866 [7], blank samples are entirely sufficient for this type of drift measurement. Herbst & Wygoda [6] also reported that the relative deviation of fluorescence intensity from the initial value, depending on the storage time, is 0% for Pyranine in a petri dish. If filter paper or a similar material as a collector were used, the relative deviation would have significantly decreased. In this case, the use of spiked samples would be necessary, according to Ahrens et al. [1].

Fig. 2

Corridors of the three wind directions

During all applications, the weather data (wind direction, wind speed, air temperature and relative humidity) were constantly recorded at a frequency of 1 Hz. Valid trials are trials with an average air temperature not exceeding 25 °C, an average relative humidity above 30% and an average wind speed between 1 and 5 m s⁻¹ in accordance with JKI-guideline 7–1.5.

Laboratory analysis

The collectors used were stored in a dark, cool room and analysed within 4 days after the trials. For the analysis, the tracer (Pyranine) was extracted from the collectors with distilled water. For this purpose, 40 mL of distilled water was filled into the collectors and shaken for 10 min on a shaking table at 65 rpm. The frequency and amplitude were chosen so that the inner walls of the collectors were completely washed around. For the analysis of Pyranine in the wash water of the collectors, the fluorometer RF-6000 (Shimadzu Duisburg, Germany) with an excitation wavelength of 405 nm and an emission wavelength of 515 nm was used [6]. The fluorometric measurement is a recognized method and is described in the JKI Guideline 7–1.5 [11] and in ISO 22866 [7]. In these drift measurements, the spray liquid is colored with the dye Pyranine, and the measurement method is only intended to detect the amount of dye, not to classify it as in a mass spectrometer.

Calculation of spray drift

Spray drift is expressed as ground sediment as a percentage of the application rate. A calibration line is used to calculate the spray drift (Eq. 1):

$${\beta }_{dep}=\frac{\left({\rho }_{smpl}-INT\right)}{{\Delta }_{calib}}*\frac{{V}_{dist}}{{A}_{colle}}$$

(1)

where βdep is the spray drift deposit [µg cm⁻²]; ρsmpl is the fluorometer reading of the sample [ −]; INT is the intercept of the calibration curve [ −]; ∆calib is the slope of the calibration curve [L µg₋₁]; Vdist is the volume of distilled water [L] and Acolle is the area of the collector to collect the spray drift [cm²]. The percentage compared to the application rate was calculated using Eq. 2:

$${\beta }_{dep\%}=\frac{{\beta }_{dep}}{TR}*100$$

(2)

where βdep% is the spray drift [%] and TR is the tracer rate [µg cm⁻²]

Values exceeding 100% mean that at that position, the application rate plus an amount caused by overspray or rebound was found. This is not due to an error in analysis but is a natural phenomenon. This mainly occurs because the application surface in these investigations was vertical and not horizontal, as with the use of a field sprayer.

Statistical analysis

The trial design is a two-factor trial design with the factor’s “nozzle” and “wind direction”. For the statistical analysis using Rstudio [17] and the packages readxl [23], lattice [19], latticeExtra [20], ggplot2 [22], ggimage [24], magick [16], gridExtra [2] and car [5].

The measured drift values are displayed in a boxplot. The boxplot shows the median (50th percentile), the 25th percentile, the 75th percentile and the extreme values of all measured values in dependency of the wind directions. The median test was used to determine differences between the wind directions at the same distance from the treated area (α = 0.05). In addition, different letters were used to indicate significant differences between the wind directions at the same distance from the treated area.

In accordance with the recommendation of the FOCUS Surface Water Working Group, the 90th percentile was used for the calculation of the basic drift values. The 90th percentile has been shown to best represent the worst-case scenario [4]. An exponential least squares regression line (best fit) was used to determine the basic drift values and the regression function was used to calculate the basic drift values for each distance. The 90th percentile was calculated from all available individual values per distance [18].

The weather conditions during the application were measured with a frequency of 1 Hz. Microsoft Excel was used to check the validity of these data and to calculate mean values and standard deviations.

Results

Chemical barrier treatment against crawling insects

Meteorological conditions during the application

Table 2 shows the weather conditions recorded when applying a chemical barrier of 50 cm split by replicates and in function of the nozzle type and target wind direction “orthogonal”, “parallel” and “shadow”. The average air temperature ranged from 12.6 °C to 23.7 °C, the average relative air humidity ranged from 40.4% to 77.6% and the average wind speed during the trials ranged between 1.0 m s⁻¹ and 3.7 m s⁻¹. These values correspond to the JKI guideline 7–1.5 [9]. Trials outside the limits were not evaluated.

Table 2 Weather conditions recorded when treating a chemical barrier of 50 cm, split by replicate and categorized by nozzle type and target wind direction

Drift values and recommended basic drift values

The following three figures show the 90th percentiles of the measured drift values as ground sediment in percentage of the application rate when applying a chemical barrier with orthogonal (Fig. 3) and parallel wind direction (Fig. 4), as well as with wind shadow (Fig. 5). For all three wind directions, the drift decreases continuously over the entire measuring area with increasing distance from the treated area. The first collector at 15 cm was declared as overspray for all three wind directions and is not included in the evaluation, as it cannot be determined whether the collectors were directly sprayed during application or whether it was a rebound from the wall.

Fig. 3

90th percentile of the measured drift values in percent of the application rate when applying a chemical barrier with orthogonal wind direction

Fig. 4

90th percentile of the measured drift values in percent of the application rate when applying a chemical barrier with parallel wind direction

Fig. 5

90th percentile of the measured values in percent of the application rate when applying a chemical barrier in the wind shadow

With orthogonal wind direction, higher drift values were observed with the brass nozzle than with the IDK 90–015 C nozzle. At a distance of 57 cm from the treated area, 0.93% drift was measured with the brass nozzle and only 0.21% drift with the IDK nozzle (Fig. 3). With parallel wind direction, the measured values were overall at a higher level than with orthogonal wind direction. At a distance of 57 cm, the drift values were 4.64% with the brass nozzle and 4.87% with the IDK 90–015 C nozzle. Over the entire measuring range, the brass nozzle produced only slightly higher drift values than the IDK 90–015 C nozzle (Fig. 4). This contrasts with the drift values when treating the area in the shadow of the wind. Without wind, the drift values of the IDK 90–015 C nozzle decrease faster with increasing distance from the treated areas than with the brass nozzle. Nevertheless, the drift values are higher than with orthogonal wind direction. Thus, in the case of wind shadow at a distance of 57 cm from the treated area, a drift of 7.72% was observed with the brass nozzle and of 4.09% with the IDK 90–015 C nozzle (Fig. 5).

Table 3 shows the recommended basic drift values when applying a chemical barrier to a building wall up to 50 cm to control crawling insects with a knapsack sprayer in different wind directions. These basic drift values were derived from the measured values using the regression line. The lowest basic drift values were observed with orthogonal wind direction, followed by parallel wind direction and treatment the area in the wind shadow. The brass nozzle showed higher basic drift values compared to the IDK 90–015 C nozzle in all cases.

Table 3 Recommended basic drift values for single application derived from the measured drift values in percent of the application rate when treating a chemical barrier based on 90th percentile

Drift mitigation measures

The trial design is a two-factor trial design with the factors “nozzle type” and “wind direction”. The analysis of the drift values showed that there was no interaction between the two factors. Therefore, the two factors can also be considered separately. Figure 6 shows the measured drift values in percent of the application rate over the three wind directions as a function of the brass nozzle and the IDK 90–015 C nozzle when applying a chemical barrier. The IDK 90–015 C nozzle shows slightly lower values than the brass nozzle, but this difference is not significant over the entire measuring area. In contrast, Fig. 7 shows the measured drift values in percent of the application rate over the two nozzles as a function of the three wind directions when applying a chemical barrier. Clearly lower drift values were observed with an orthogonal wind direction than with parallel wind directions and with treating in the shadow. This difference is significant over the entire measuring area.

Fig. 6

Measured drift values in percent of the application rate when treating a chemical barrier as a function of nozzle type (Median.Test, α = 0.05)

Fig. 7

Measured drift values in percent of the application rate when treating a chemical barrier as a function of the wind direction (Median.Test, α = 0.05)

Treating an entire building wall against flying insects

Meteorological conditions during the application

Table 4 shows the weather conditions recorded when treating a building wall split by replicates and in function of the nozzle type and target wind directions (orthogonal, parallel and shadow). The average air temperature ranged from 7.3 °C to 25.0 °C. The average relative air humidity ranged from 39.7% to 63.6% and the average wind speed during the trials ranged between 1.6 m s⁻¹ and 4.3 m s⁻¹. These values correspond to the JKI guideline 7–1.5 [9]. Trials outside the limits were not evaluated, with the exception of one trial repetition, during which an average air temperature of 25.5 °C was measured during an application in the shadow. However, the drift values did not show any abnormalities, so these data were also included in the data set.

Table 4 Weather conditions recorded when treating a building wall, split by replicate and categorized by nozzle type and target wind direction

Measured drift values and recommended basic drift values

Figures 8 and 9 show the 90th percentiles of the measured drift values in percent of the application rate when treating a building wall at orthogonal and parallel wind direction. For both wind directions and with both nozzles, the drift values decrease with increasing distance from the building wall. The first collector at 15 cm was declared as overspray for all three wind directions and is not included in the evaluation, as it cannot be determined whether the collectors were directly sprayed during application or whether it was a rebound from the wall. In addition, it can be seen that for both wind directions, the drift values of the IDK 90–015 C nozzle are at close range higher than the drift values of the brass nozzle, but in greater distance the drift values of the IDK 90–015 C nozzle decrease faster than the drift values of the brass nozzle. This effect is more pronounced with orthogonal wind direction than with parallel wind direction. The drift values for parallel wind direction are many times higher than the drift values for orthogonal wind direction. Thus, at a distance of 57 cm from the treated area and with parallel wind direction, a drift of 39.88% was measured with the brass nozzle and a drift of 51.82% with the IDK 90–015 C (Fig. 8). With orthogonal wind direction, a drift of 12.67% was measured with the brass nozzle and a drift of 14.65% with the IDK 90–015 C at this distance (Fig. 9).

Fig. 8

90th percentile of the measured drift values in percent of the application rate when treating a building wall at orthogonal wind direction

Fig. 9

90th percentile of the measured drift values in percent of the application rate when treating a building wall at parallel wind direction

When treating a building wall in the wind shadow, it was also observed that the drift values decreased with increasing distance from the treated area and that the drift values of the IDK 90–015 C nozzle were higher than the drift values of the brass nozzle in the close range of the treated area, but decreased faster than the drift values of the brass nozzle.

Another effect, however, which was not observed with the other wind directions nor during the application of the chemical barrier, is the vortex effect at the edge of the building wall. Due to the fact that the user was standing in the wind shadow of the container, turbulence occurred at the edge of the container. Figures 10 and 11 show the 90th percentiles of the measured drift values in percent of the application rate when treating a building wall in the wind shadow, categorized into data for the entire wall, data for edge only and data for the wall without data generated at the edge. For both tested nozzles, the drift values without taking the edge into account are significantly lower than with the vortex effect at the edges or when only the edge is taken into account. For the IDK 90–015 C nozzle, the drift values for the wall without data from the edges are two to three times lower than the drift values of the entire wall and, in some cases, four times lower than the drift values considered only the edge. For the brass nozzle, at a distance of 57 cm from the treated area, a drift of 66.34% was observed using data for the edge only, 55.29% for the entire wall, and 39.16% for the wall without the edge (Fig. 10). In contrast, for the IDK 90–015 C nozzle, at a distance of 57 cm from the treated area, a drift of 103.23% was observed for the edge only, 63.34% for the entire wall, and 25.51% for the wall without the edge (Fig. 11).

Fig. 10

90th percentile of the measured values of the drift in percent of the application rate when treating a building wall in the wind shadow with the brass nozzle

Fig. 11

90th percentile of the measured values of the drift in percent of the application rate when treating a building wall in the wind shadow with the IDK 90-015 C nozzle

Table 5 shows the recommended basic drift values for a treatment of a building wall at three different wind directions based on the 90th percentile. These basic drift values were derived from the regression lines of the measured drift values. As with the application of a chemical barrier, the lowest basic drift values were also observed for the treatment of the building wall with orthogonal wind direction, followed by parallel wind direction and treatment in the wind shadow. For the treatment in the wind shadow, the drift values without edge effect were taken into account, as the main aim of the study was to measure the variances in drift depending on the wind direction.

Table 5 Recommended basic drift values for single application derived from the measured drift values in percent of the application rate when treating a building wall based on 90th percentile

Drift mitigation measures

In contrast to the application of a chemical barrier, a significant effect of the nozzles used can be observed when treating an entire building wall. Figure 12 shows the measured drift values in percent of the application rate when treating a building wall as a function of the nozzle type. At a distance of 29 cm to the treatment area, the drift values when using the IDK 90–015 C nozzle is significantly higher than the drift values when using the brass nozzle. However, the drift values when using the IDK 90–015 C nozzle decrease faster than the drift values when using the brass nozzle. This resulted in significant lower drift values when using the IDK 90–015 C nozzle than when using the brass nozzle from a distance of 71 cm and more to the treatment area. Just as during the application of a chemical barrier, a significant effect of the wind direction can also be seen when treating the building wall. With orthogonal wind direction, significantly lower drift values were measured than with parallel wind direction or treatment of the building wall in the wind shadow (Fig. 13).

Fig. 12

Measured drift values in percent of the application rate when treating a building wall as a function of the nozzle type (Median.Test, α = 0.05)

Fig. 13

Measured drift values in percent of the application rate when treating a building wall as a function of the wind direction (Median.Test, α = 0.05)

Discussion and recommendation

Comparison of derived basic drift values to the parameter “Fraction emitted to soil during outdoor wall spray application due to deposition (F_{spray, deposition})” in appropriate emission models presented in OECD ESD PT18 No. 18 (2008).

These data are the first to describe biocide deposition on buildings after application. This makes it difficult to discuss and compare the data with data from the literature. Only in 2008, OECD published an Emission Scenario Document for insecticides, acaricides and other arthropod control products for buildinghold and professional use [15]. This document inter alia provides harmonized models to assess environmental emission and exposure due to insecticide spray applications in outdoor areas. A number of standard parameters (e.g., fractions emitted to soil during application and) are assumed but not measured. Both in the scenario to model treatments against flying insects as well as in the scenario to model treatments against crawling insects by spray applications on building walls a default fraction of product subject to deposition of 10% is used to map part of the exposure of the surrounding soil compartment. In addition, the parameter “Fraction emitted to soil during outdoor wall spray application due to deposition (F_{spray, deposition})” is based on the former OECD ESD for masonry preservatives [13] as no other data is currently available. The fractions released due to application on walls (including deposition) will reach the soil adjacent to the building; for the assessment of predicted environment concentrations (PEC) it is proposed to define a receiving soil area defined by perimeter of building and 50 cm distance from the treated walls. The default value of 50 cm from the emission scenario document is compared with the measured values at 57 cm from the treated area. When applying a chemical barrier, with both nozzles and at the three wind directions, the default fraction emitted to soil due to deposition of 0.1 was considered appropriate at a distance of 57 cm to the treatment area. For the treatment of the entire building wall, this is not the case as all derived basic drift values at 57 cm distance are higher than the OECD default value, independent from wind direction or nozzle type.

Drift reduction measures

When treating an entire building wall or a part of it with a chemical barrier, the wind direction has a significant effect on spray sedimentation. During treatment with orthogonal wind direction, the drift values are significantly lower than with parallel wind direction and when treating the building wall in the shadow of the wind. The drift values are lower with orthogonal wind direction because the droplets are "pressed" against the wall and thus less prone to drift. This is irrespective of the nozzles. If the wind direction was parallel to the building wall or if the wall was treated in the shadow of the wind, the drops dispersed in front of the treated area. This effect is even stronger when treating an entire building wall than when applying only a chemical barrier of 50 cm height, as the treated area is larger. Another effect, which was only observed when treating the entire building wall, is the effect of vortex at the edge of the building wall when treating in the wind shadow. A much higher drift was observed at the two lateral edges than in the middle of the treated area. Therefore, the results were also presented separately in the three areas "entire wall", "edge only" and "wall without edge". It is very difficult to draw a conclusion to minimize emissions to the environment related to wind direction. Usually, not only one wall is treated, but all walls of the building. Thus, all wind directions will be relevant during an application process. A drift mitigation measure based on the wind direction would therefore not be practically feasible, as it would require the treatments to take place based on changes in wind directions.

However, the results show that when the entire building wall is treated, the choice of nozzle has an influence on the drift reduction potential. The choice of a nozzle with a droplet size distribution from very coarse to medium, such as the IDK 90 015 C nozzle, leads to lower emissions into the environment from a distance of 71 cm.

Choice of nozzle type and exposure in relation to distance to the treatment area

The results for the treatment of the entire building wall show that evaluating the drift of a sprayer or nozzle based on only one drift value at one distance can lead to misinterpretations. Both with orthogonal and parallel wind direction, a higher drift was measured with the IDK 90-015 C nozzle than with the brass nozzle at a distance of 57 cm from the treated entire building wall. However, to get the full picture, it needs to be taken into account that the drift decreased faster with increasing distance for the IDK 90-015 C nozzle than for the brass nozzle, due the different droplet size of these two nozzles. In greater distance to the treated area, the drift from the IDK 90-015 C is, therefore, lower.

This effect can be explained by the design of the different nozzle types. A hollow cone nozzle, such as the brass nozzle, has a swirl plate with one or more tangential of helical slots or hole. Liquid is forced through this swirl plate into a swirl chamber. An air core is formed as the liquid passes with a high rotational velocity from the swirl chamber through a circular size due owing to the tangential and axial components of velocity [12]. An injector flat spray nozzle, such as the IDK 90-015 C, consist of an injector insert with a dosing hole at the top and a mouthpiece at the bottom for distributing the liquid. A mixing chamber is located in between. The nozzles draw in air through the injector according to the Venturi principle (via holes on the side or front). Air and liquid are mixed in the mixing chamber. The liquid/air mixture leaves the nozzle opening at a greatly reduced pressure, which is considerably wider than with normal flat spray nozzles of the same capacity. The result is relatively coarse, very drift-stable droplets with high energy [14].

The laboratory analysis of these two nozzles showed that the spray angle is very similar, but that the distribution of the droplets is very different. While the brass nozzle has a VMD of 208.2 µm at 2.0 bar and is thus classified as "Fine", the VMD of the IDK 90–015 C nozzle is 561.9 µm at 2.0 bar and is classified as "Extremely Coarse" (Table 1). These properties are reflected in the drift tests as follows: When applying a chemical barrier of 50 cm to control crawling insects, the lowest drift was measured with the IDK 90–015 C in all three wind directions. The difference was most pronounced when the wind direction was orthogonal. When treating an entire building wall for flying insect control, the drift decreased faster with the IDK 90–015 C nozzle than with the brass nozzle but the IDK 90–015 C had higher drift values close to the treated wall. This pattern was observed for all three wind directions and especially for the applications in the wind shadow, when the wind theoretically has no influence on the drift and only the nozzle properties were influencing the drift. In this case, the larger drops of the IDK 90–015 C nozzle fall faster to the ground due to gravity and thus increase the drift in this area. Another possibility for higher values of the IDK 90–015 C close to the treated area could also be a rebound effect. Large droplets have a greater kinetic energy and can, therefore, rebound more strongly than smaller droplets. The smaller drops of the brass nozzle drift further and thus increase the drift at the greater distance from the treated area.

Thus, although the drift close to the treated area is higher with the IDK 90–015 C nozzle, the total contaminated area is smaller with the IDK 90–015 C nozzle than with the brass nozzle. If only one drift value at one distance would be regarded, this would not be taken into account. This is relevant for the comparison of different techniques as well as for the environmental exposure assessment.

Further development of guidance

The JKI guideline 7–1.5 for measuring direct drift when applying liquid plant protection products in the field was used as a basis for the measurements presented in this study [9]. It was adapted to the specific scenario of treating a building wall by reducing the measuring distance from 50 m to 1.83 m for foundation application and 1.42 m for building wall application. Due to the differences in the experimental set-up, the drift values derived in this study are not comparable to drift values derived for nozzles used for the application for plant protection products using the original set-up. It is recommended to develop an own guideline for drift measurements in this biocidal setting, including further investigations on rebound effects. The recommended drift values in these studies are the first values that have not been used before. These values can be used to derive the first risk assessments required for product approval. Similarly, knowledge of the parameters that influence drift can serve as a basis for defining general requirements for machines in product approvals to reduce environmental risks.

In summary, the study shows for the first time that the detailed knowledge on drift behaviour for specific settings and techniques can help in reducing the environmental pollution with biocides in non-target areas and also might increase the efficient use of biocidal products by more targeted applications.