2014 Volume 39 Issue 1 Pages 14-21
Estimates and risk assessments of spray drift have mainly concentrated on drift from the field into horizontal areas such as neighboring fields or water bodies (e.g. Ganzelmeier et al. 19951)). However, experiments simulating pesticide drift have shown that some common berry-producing hedgerow tree species are very sensitive to herbicides. Thus, several studies found that sulfonylurea herbicides reduce growth and fruit yield of cherry trees at dosages as low as 1% of the label rate,2–4) and Kjær et al. (2006)5,6) showed that drifting metsulfuron methyl may reduce the growth and berry production of hawthorns, both in the year of exposure and in the following year. Since the berries of e.g. hawthorn, are not uniformly distributed over the height of the trees, pesticide deposition at different heights may have different effects. Therefore, knowledge of the vertical distribution of pesticides drifting into hedgerows is very important.
Previous measurements of pesticide spray drift have concentrated on soil deposition at varying distances from the spray boom. Most studies have only included one run by the tractor, i.e., no integration of the total drift from spraying an entire field. The main conclusion from these measurements was that most of the pesticide is deposited within the nearest 5 m.1,7–9) The few studies that have compared cumulative deposition from one and ten runs show that integration of contributions from ten swaths increases spray drift with a factor of 3–10 for horizontal deposition measures.10–12) The accumulated deposition at a specific position downwind was calculated from the measured deposition from a single swath at varying distances from the sprayer. A previous study that measured pesticide deposition in hedgerows found that between 2 and 20% of the dosage applied in the swath closest to the hedgerow drifted into the hedgerow.13) This study showed that at heights up to 2 m above ground, 15–20% of the dosage applied in a single swath 6 m from a hedgerow at a boom height of 1.4 m was deposited on drinking straw samplers placed at a distance of 0–4 from the hedgerow, i.e., 2–6 m from the sprayer, which was mounted with Albuz APG 110V nozzles. Wind speed at boom height was measured to be 3.4 to 4.7 m/s. Furthermore, hedgerows have been shown to reduce spray drift to neighboring fields by up to 90% compared to fields without hedgerows.14–18) At least some of that 90% is likely to be deposited in the hedgerow. These results support the importance of including contributions from all swaths when estimating how pesticides are deposited in hedgerows.
It is the aim of this study to describe the importance of height, distance and weather conditions for the spray drift to hedgerows. In addition, some settings of the spray equipment were varied in order to qualify the discussion of other necessary variables for the assessment of spray drift. The outcome of the experiments may be used for assessing the efficiency of mitigation measures in terms of buffer zones for potentially exposed/vulnerable organisms living in different parts of the hedgerows.
A series of seven spray experiments was carried out in order to determine the pesticide deposition in hedgerows due to spray drift from multiple spray swaths. These experiments were conducted so that deposition at different heights of the hedgerow could be determined in order to estimate the effect of an unsprayed buffer zone on spray drift.
The experiments took place in a spring barley field surrounded by hawthorn hedgerows on three sides (west, north and east of the field). In each experiment, spray drift from five swaths with increasing distance to a hawthorn hedgerow was measured individually, and the experiment was repeated on seven occasions (April 2005, May 2005, June 2005, August 2005, September 2005 and twice in June 2010) in the hedgerow approximately perpendicular to the actual wind direction on the day of spraying. The deviations in the seven trials were 3–9°, 3–20°, 6–26°, 23–37°, 19–40°, 23–37° and 11–24°, respectively. The increased drift distances at large deviations from 90° were not taken into account.
1. Spray applicationSpraying was performed with a conventional tractor-mounted sprayer in five tracks parallel to the hedgerow. The employed spray nozzles were Hardi 4110-16 (Hardi, Denmark; comparable to the Hardi ISO 025 or 03 with respect to liquid delivery rate, BCPC class “Fine”) in 2005 and TeeJet AI 110-04 (TeeJet, Illinois, USA; BCPC class “Coarse” to “Very Coarse”) in 2010. The spray boom was 12 m wide with a nozzle distance and boom height of 50 cm. Thus, distances between the near end of the spray boom and the hedgerow were 0, 12, 24, 36 and 48 m for the different swaths, except for June and August 2005, where the distances between hedgerow and spray boom were 4 m longer due to a 4 m wide dirt road running along the hedgerow. The tractor had a driving speed of 6.4 km/h in 2005 and 7 km/h in 2010 (cf. Table 1).
| Parameter, unit | Trial | |||||||
|---|---|---|---|---|---|---|---|---|
| April 05 | May 05 | June 05 | Aug. 05 | Sept. 05 | June 10 | June 10 | ||
| Environmental Conditions | Crop density | Sparse | Sparse | Dense | Stubbles | Stubbles | Dense | Dense |
| Crop height, cm | 2 | 5 | 20 | 10 | 10 | 20 | 20 | |
| Hedgerow orientation | West | North | East | East | West | East | East | |
| Wind speed, m/s | 5.5–6.2 | 2.0–3.6 | 2.9–3.8 | 3.7–4.5 | 2.4–3.3 | 4.6–5.8 | 4.8–5.6 | |
| Wind direction, °C | 114–126 | 197–220 | 271–303 | 260–274 | 136–157 | 260–274 | 273–286 | |
| u*, m/sb) | 0.51–0.61 | 0.19–0.28 | 0.39–0.45 | 0.34–0.53 | 0.29–0.58 | 0.45–0.62 | 0.51–0.61 | |
| L, mc) | −57–−88 | −5–−21 | −58–−88 | −21–−87 | −24–−107 | 244 | −116–−196 | |
| Heat flux, W/m2 | 206–255 | 51–130 | 74–120 | 93–205 | 91–164 | 44–112 | 85–144 | |
| Temperature, °C | 10–11 | 11–12 | 17–18 | 17–18 | 22–24 | 12–15 | 15–16 | |
| Glob. Rad.d) W/m2 | 640–700 | 470–440 | 600–750 | 410–490 | 420–480 | 374–895 | 602–920 | |
| Absolutee) hum., g/kg | 3.6–4.0 | 4.6–4.9 | 8.6–9.1 | 6.4–7.3 | 9.0–10.4 | 6.6–6.8 | 6.6–6.7 | |
| Spraying equipment settings | Nozzle typef) | FF | FF | FF | FF | FF | AI | AI |
| Pressure, mPa | 0.55 | 0.55 | 0.3 | 0.3 | 0.3 | 0.3 | 0.3 | |
| Flow rate, L/min | 1.6 | 1.6 | 1.1 | 1.1 | 1.1 | 1.6 | 1.6 | |
| Speed, km/h | 7 | 7 | 7 | 7 | 7 | 6.4 | 6.4 | |
| Application rate, L/ha | 300 | 300 | 200 | 200 | 200 | 300 | 300 | |
| Tank concentration, g/L | 1.49 | 2.23 | 1.63 | 1.99 | 1.83 | 2.23 | 1.93 | |
a) Relatively thin cloud cover b) The parameter u* is the friction velocity, a measure of air turbulence c) L (Monin–Obukhov number) is a parameter of atmospheric stability d) “Glob.rad.” is the global radiation e) Calculated on basis of relative humidity, temperature and air pressure f) FF=Hardi flat fan 4110-16 nozzles and AI=TeeJet Air induction 110-04
In order to determine spray drift, sodium fluorescein was added to the spray liquid as a tracer dye at concentrations of 1.49, 2.23, 1.63, 1.99, 1.83, 1.93 and 2.23 g/L, respectively. On one occasion (May 2005), the herbicide metsulfuron methyl was applied together with the dye marker sodium fluorescein in order to establish the relationship between herbicide and dye deposition. The concentration of metsulfuron methyl in the spray liquid was 0.02 g/L. At all other spraying events, only the dye was applied, since the dye is both cheaper and easier to analyze.
Meteorological conditions were registered (Table 1) by measuring relative humidity (Theodor Friederichs Humidity Sensor 3030 (Schenefeld, Germany) at 1.3 m height), temperature, wind speed, wind direction, turbulence, heat flux (Metek USA-1 Ultrasonic Anemometer (Elmshorn, Germany) at 4 m height) and global radiation (Soldata SPC80 Pyranometer, Silkeborg, Denmark). The equipment was mounted on a 6 m high mast placed in the center of the field. Absolute humidity was calculated on the basis of relative humidity, temperature and air pressure, the latter obtained from Deutscher Wetterdienst Offenbach (DWD) at http://www.wetter3.de/Archiv/archiv_dwd.html, with an accuracy of 1 hPa.
2. Characterization of nozzlesThe droplet size distributions of the Hardi 4110-16 flat fan nozzle at 0.30 mPa and 0.55 hPar and the TeeJet air induction nozzle 110-04 at 0.3 mPa were experimentally determined, using the Phase Doppler Particle Analyzers (PDPA) laser-based measurement setup and protocol described by Nuyttens et al. (2007, 2009).19,20) Droplet size and 1D-velocity measurements were performed at 0.05 m below the spray nozzle by scanning a rectangular pattern at a constant speed to achieve a complete scan of the spray plume close to the nozzle exit. All measurements were made using tap water at a temperature of 20°C. Environmental conditions were kept constant at a temperature of 20°C and a relative humidity of 60–70%. Each scan yielded data for at least 20,000 droplets.
3. Spray deposition measuresFor each of the five swaths in a single experiment, herbicide and dye were collected on leaves from the hedgerow and on hair curlers mounted uprightly on five masts placed at 10 m intervals right next to the hedgerow trees (Fig. 1). First the 12 m wide swath closest to the hedgerow was sprayed; thereafter the exposed curlers and leaves were sampled. Then the second swath, placed 12 m farther way from the hedgerow, was sprayed and curlers and leaves collected, and so on until all five swaths had been sprayed and all samples collected. The selection of sampling methods was made to achieve both standardized measures (curlers) and ecologically relevant measures (leaves). Due to the different analysis techniques (see below), herbicide and dye were analyzed on separate samples.
3.1. Leaf samplesAt the time of the first spraying in April 2005, leaves had not yet emerged. At all other spray events, each leaf sample consisted of 5–10 leaves from a single branch on the “surface” of the hedgerow facing the sprayed field. Samples were taken in the immediate vicinity of the masts carrying the curlers, at app. 0.5, 1 and 2 m above ground with five replicates per height and swath (Figs. 1 and 2). Prior to spraying, all branches selected for sampling were covered with plastic bags (Fig. 2). Just before the spraying of a given swath, the relevant plastic bags were removed in order to ensure that each leaf sample was only exposed to spray drift from that particular spray swath. Immediately upon exposure, the leaves were sampled, placed in plastic cylinders containing extraction fluid (0.1 M Na2HPO4 buffer or deionized water) and stored in darkness to prevent photo degradation of the dye. In the laboratory, the total one-sided leaf area was determined by means of an LI-3100 Area Meter (LI-Cor. Inc., Nebraska, USA).
Prior to the spraying of a swath, the five masts (replicates) placed in the hedgerow were equipped with two hair curlers (commercial plastic hair curlers of the brand M-cosmetics, diameter 2 cm, length 6 cm) at each height (0.5, 1, 2 and 4 m) in order to measure dye marker deposition (Figs. 1 and 2). In May 2005, when herbicide was applied, one more curler was added at each height on the masts placed in the hedgerow in order to sample herbicide deposition at spraying 0, 24 and 48 m from the hedgerow. Upon spraying, curlers were treated as described for leaves. Results were expressed as the amount of sodium fluorescein per curler as all curlers have the same collection area.
3.3. Analysis of sodium fluoresceinSodium fluorescein deposition was measured by fluorescence spectrophotometry upon extraction in a 0.1 M Na2HPO4 buffer adjusted to pH 6.8 with NaH2PO4. Fluorescein is excited at 492 nm and was detected at 513 nm by a fluorescence HPLC monitor (Shimadzu RF-551, Shimadzu, Kyoto, Japan; detection limit 0.01 µg/L). In addition to leaf and curler samples, a sample from the spray tank was analyzed from every spraying occasion. Prior to the experiments, the potential decay of sodium fluorescein was estimated and no decay could be detected during two days of storage at room temperature, which indicates that no measurable decay took place between sampling and analysis. Since the period of time from when a swath was sprayed until all samples were placed in darkness did not exceed 20 min, the risk of photo degradation in the field is considered negligible.
3.4. Analysis of metsulfuron methylMetsulfuron methyl was extracted from curlers with deionized water. Chlorsulfuron was added to the water extract as a surrogate standard, since isotope labeled metsulfuron was not available at the time of the experiments. The water extract was further concentrated on solid phase extraction (SPE) cartridges (Oasis HLB, 200 mg, Waters, Milford USA). Detection and quantification of the target compounds were performed with liquid chromatography-tandem mass spectrometry (LC-MS-MS) equipped with electrospray ionization (ESI) operating in positive ionization mode. The analytes were separated on a Hypersil C18 column 2.1×25 cm, 5 µm particle size (Phenomenex) with 5 mM ammonium acetate buffer and methanol as mobile phase A and B, respectively. The mass spectrometer was operated in multiple reaction monitoring (MRM) for the following precursor-product transition ions: 382/167 for metsulfuron methyl and 358/167 for chlorsulfuron.
Leaf areas (one-sided) were measured by means of a LI-3100 Area Meter (LI-Cor. Inc., Nebraska, USA) before the leaves were cut and crushed in liquid nitrogen. The samples were spiked with chlorsulfuron and mixed with a 25% aqueous solution of methanol (pH adjusted to 12 with sodium hydroxide). This mixture was extracted ultrasonically for 15 min. The resulting extract was filtered through nylon mesh (0.42 and 0.2 µm) and analysed by LC-MS-MS. The samples were extracted and analyzed in batches together with a procedural blank. The target compounds were not detected in any of the blank samples. The detection limit of the analytical method (MDL) was defined as those concentrations of the analytes needed to produce a signal-to-noise ratio (S/N) of 3 : 1.
The detection limit for metsulfuron methyl was 0.8 and 3 ng per sample for curlers and leaves, respectively.
3.5. Buffer zone calculationsUnsprayed buffer zones have been used as a tool to reduce pesticide drift to surrounding areas, especially to water bodies. In the present study, drift is expressed as estimated area-based deposition on leaves (µg/cm2 brought to the scale of L/ha) as a percentage of the area-based application rate in the field (L/ha). The actual deposition given in µL/curler was multiplied by the conversion factor from curler deposition to leaf deposition per area (Table 3) and divided by the actual tank concentration. This gives a deposition equivalent to µL spray liquid/cm2.
The accumulated deposition was calculated as
 |
For calculation of the importance of unsprayed buffer zones for the spray deposition in adjacent habitats (hedgerows), the contributions from three swaths were summed. As only five swaths were established in the experiments, the buffer zones in the calculations have a width of 0, 1 or 2 spray swaths (0, 12 and 24 m for five spray occasions; 4, 16 and 28 m for the two occasions with a dirt road between the field and hedgerow), since wider buffer zones would lead to integration of fewer than three swaths. Consequently, the summation for the scenario without a buffer zone consisted of swaths 1, 2 and 3; in the case of a 12 m buffer zone the summation involved deposition from swaths 2, 3 and 4. The effect of a 24 m buffer zone was calculated on basis of deposition measures from spray swaths 3, 4 and 5.
These calculations were made separately for each trial because the conversion factors differed between experiments. Since there was no conversion factor for the first experiment as the leaves had not yet emerged, the average conversion factor of the other six trials (0.02) was used.
4. Statistical methodsLinear regression analysis was used to estimate the relationship between measured herbicide deposition and tracer deposition as well as the relationship between tracer deposition on artificial targets (hair curlers) and hawthorn leaves (trial repeated seven times). In each trial, runs with five swaths of increasing distance to sprayer were conducted with sampling five replicates for each of the four sampling heights.
The spray deposition data were analyzed using a multiple model, i.e., a regression model with several explanatory variables (generalized linear models; GLM). The initial statistical model was as follows:
 |
Before estimating the statistical parameters (βx), the response variable “deposition” was transformed by the natural logarithmic function, resulting in Gaussian distributed residuals and a multiplicative model on the non-transformed scale of the response variable. To avoid the problem of multi-collinearity in regression analysis, Pearson correlation coefficients were calculated between each pair of continuous explanatory variables. No large correlations (Pearson’s correlation coefficients between −0.3 and 0.3) were found, and consequently, all explanatory variables could be included in the primary statistical model. Any non-significant terms were excluded and parameters were re-estimated. As measures of initial droplet size distribution V100 (percentage droplets smaller than 100 µm), V50 (percentage droplets smaller than 50 µm) and VMD (median droplet diameter) were tested.
The characteristics of the two nozzle types are summarized in Table 2. As expected, the Hardi flat fan nozzle produced smaller droplets than the TeeJet air induction nozzle at equal working pressure (0.30 mPa). Furthermore, increasing the working pressure to 0.55 mPa caused the Hardi nozzle to produce slightly smaller droplets.
| Hardi 4110-16, 0.3 mPa | Hardi 4110-16, 0.55 mPa | TeeJet AI 110-04, 0.3 mPa | |
|---|---|---|---|
| V50 (%) | 1.00±0.06 | 0.87±0.13 | 0.029±0.001 |
| V100 (%) | 8.99±0.73 | 9.99±1.79 | 1.80±0.19 |
| VMD (µm) | 213.2±8.8 | 198.1±16.4 | 439.4±6.5 |
The comparison of deposition measures of the herbicide metsulfuron methyl and the dye tracer fluorescein showed that tracer deposition and herbicide deposition measured in separate samples were well correlated (metsulfuron (µg/curler)=0.01×tracer (µg/curler) N=60, R2=0.71). Since there was a factor of 100 difference in concentration between metsulfuron methyl and fluorescein in the spray liquid, the established relationship between pesticide and tracer means that the rate of deposition on curlers is similar for the two substances. Furthermore, it was found that tracer deposition on hair curlers was significantly correlated to tracer deposition on leaf surfaces in the hawthorn hedgerow (Table 3). However, the relation differs between hedgerows and between days.
Fig. 3 demonstrates that deposition was highly variable between single trials, even though some trials (6 and 7) were conducted in the same place, on the same day and with the same equipment. Further, the subfigures show that deposition depended on distance to sprayer and height in hedgerow. Analysis of the influence of equipment settings and weather conditions showed that initial droplet size distribution, distance of the spray swath to the hedgerow, height of the target in the hedgerow, absolute humidity and wind speed at the time of spraying had significant effects on spray deposition in the hedgerow (Table 4). The resulting statistical model, leaving out temperature and intercept, which turned out not to significantly affect spray drift, and back-transforming the log-transformed deposition data, is:
| Trial | α±S.E. | N | R2 | F | p |
|---|---|---|---|---|---|
| 2 | 0.0307±0.00441 | 69 | 0.42 | 48.29 | <0.0001 |
| 3 | 0.0200±0.00110 | 73 | 0.82 | 335.95 | <0.0001 |
| 4 | 0.0133±0.00086 | 73 | 0.77 | 240.55 | <0.0001 |
| 5 | 0.0186±0.00189 | 69 | 0.59 | 96.91 | <0.0001 |
| 6 | 0.0152±0.00182 | 75 | 0.49 | 69.77 | <0.0001 |
| 7 | 0.0228±0.00130 | 71 | 0.81 | 309.28 | <0.0001 |
| Effect (parameter) | Parameter estimate | F | N | DF | p | R2 |
|---|---|---|---|---|---|---|
| Model | 747.43 | 801 | 5 | <0.0001 | 0.824 | |
| Wind speed (β2) | −1.44 | 31.46 | <0.0001 | |||
| Absolute humidity (β3) | 0.279 | 124.64 | <0.0001 | |||
| V100 (β4) | −0.0636 | 80.12 | <0.0001 | |||
| Height in hedgerow (β5) | −0.270 | 112.35 | <0.0001 | |||
| Distance to hedgerow (β6) | −0.0472 | 679.91 | <0.0001 |
 |
The bias-correction factor needed when back-transforming (e1.77) is derived according to Ferguson (1986).21) All measures of initial droplet size distribution showed the same trend, with V100 giving a slightly better description (R2) than V50 and VMD. The resulting statistical model described 82% of the observed variation. The F-statistic of the analysis of variance gives the importance of the influential factors. The primary single impact factor was “Distance to hedgerow” followed by “Absolute humidity,” “Height in hedgerow,” V100 and wind speed.
The calculations of accumulated deposition from three spray swaths show that for the Hardi flat fan nozzles (trials 1–5) without a buffer zone, the deposition at a height equivalent to the height of the spray boom varied between 1.5 and 8% of the application rate (solid line in Figs. 4 A–E, height 0.5 m). At a height of 4 m, this was reduced to a deposition between 0.36 and 0.77%.
By introducing a buffer zone of 12 or 24 m, deposition as a function of height was changed so that there were only small differences in deposition with height for the different buffer zones (broken lines in Figs. 4 A–E), which demonstrates that the effect of buffer zones was largest in the lower parts of the hedgerow. The deposition at 4 m height was nearly independent of the establishment of buffer zones as long as the same field area was sprayed. On average (±SEM), depositions at the heights of 0.5, 1, 2 and 4 m were reduced by 72±4, 66±4, 47±8 and 0.15±14%, respectively, by inserting a 12 m wide unsprayed buffer zone between the field and hedgerow. For a 24 m wide buffer zone, the corresponding reductions were 88±3, 85±4, 73±8 and 41±14%.
The present study documents that the level of spray drift is different for different heights in a hedgerow adjacent to a sprayed field and that introduction of an unsprayed buffer zone will have different effects at different heights. This is important because different organisms utilize different parts of the hedgerow for feeding and nesting. Our results are in agreement with Parkin and Merritt (1988),22) who measured deposition from Lurmark 04-F110 nozzles at heights between 0 and 12 m and found that spray drift from short distances decreased rapidly with height, whereas at longer distances from the sprayer deposition only decreased slightly with height. Weisser et al. 2002,23) who studied the deposition on plant surfaces in a hedgerow, also found decreased deposition from the bottom of the hedgerow up to a 2 m height for both a low-drift nozzle AI 110 025 and the conventional flat fan nozzle (XR 110 03). In one experiment, Longley et al. (1997)13) found only very small differences between 1, 1.5 and 2 m heights when combining data from distances of 2, 4 and 6 m from the spray boom. However, our results demonstrate that drift from longer distances contributes more to the total drift at the top of the hedgerow than to the total drift at the lower part of the hedgerow (Fig. 4). In addition, boom height was 1.4 m in the Longley et al. experiment, i.e., not much lower than the highest points of measuring. In another experiment, Longley and Sotherton (1997)24) found maximum deposition at the height of the spray boom and diminishing values upward.
Taking into account the fact that fluorescein and herbicide were measured on separate curler samples, the former proved a good predictor of the latter, which demonstrates the applicability of dye markers in spray drift experiments. Fluorescein deposition on hair curlers also was a good measure for herbicide deposition on leaf surfaces in hawthorn hedgerows. However, the relation between deposition per curler and deposition per leaf area differed between hedgerows and between days, probably reflecting variation in factors such as leaf development and size, hedgerow porosity and weather conditions.25,26) However, we have no indications of how collecting efficiency of the curlers depends on meteorological and technical factors.
It is interesting not only how deposition varies with environmental conditions and the setting of the spraying equipment but also whether the observed levels of deposition can evoke detrimental effects on non-target organisms. The effects of drifting insecticides on insects in hedgerows have been studied by Longley and Sotherton (1997),24) who found that a single swath of 10 m caused a 25% mortality rate for second instar Spodoptera littoralis, Boisd. larvae. Similarly, Cilgi and Jepson (1995)27) found an expected mortality rate of 75% from deltamethrin for fourth instar larvae of Pieris brassicae L. (Lepidoptera, Pieridae) at levels comparable to 0.01–2% of field rate, i.e., at exposure levels that according to our results are very likely to occur in hedgerows. Thus, insecticide effects from drift into hedgerows may occur.
For plants, several studies have assessed the impact of sulfonylurea herbicides on berry.producing trees, i.g., bird cherries, sweet cherry and hawthorn.2,3,6) Al-Khatib et al. (1992)2) found a 13–60% effect for different end points at a dose of 3% of the field rate. Kjær et al. (2006)6) demonstrated nearly 100% mortality at exposure levels observed at spray drift under normal spray conditions (2.5–5% of maximum recommended field rate) and Kjær et al. (2006)5) found that reproduction was still affected a year after exposure at an exposure of 10% of the maximum recommended field rate. Consequently, the drift rates measured in the present experiment are also likely to cause effects on hedgerow plants, at least for some herbicides.
In order to assess the risk of effects of spray drift on hedgerow plants or other organisms living in the hedgerows, it is necessary to be able to predict the spray deposition under the conditions most relevant for the area of concern and to be able to differentiate between organisms living at different heights in the hedgerow. Unsprayed buffer zones are used as mitigating measures to reduce deposition outside the cropped area.8) The background for this practice is spray drift assessments developed for water bodies and measured by means of targets placed horizontally at a low height. The fact that unsprayed buffer zones will have a large effect on spray drift at the bottom of the hedgerow, but only small or even negligible effects on drift at the upper part, strongly indicates that buffer zones are a useful mitigation measure for organisms living in the lower parts but not necessarily for those living at the top of the hedgerow. Therefore, protection of hedgerows and other habitats with a significant vertical component may require other measures such as drift-reducing nozzles and careful choice of spraying conditions. For hedgerow trees, this need is further heightened by the fact that many species carry the majority of their flowers and berries in the top part.
We found that both weather conditions and initial droplet size distribution as affected by the type and pressure of nozzles were important for the size of the deposition in the hedgerow. Thus, in our experiments the primary impact factor was distance to hedgerow followed by absolute humidity, height in hedgerow, percentage of droplets smaller than 100 µm and wind speed. Another study24) has also shown that weather conditions such as wind speed are important. Arvidsson et al. (2011)28) measured both horizontal fallout and airborne drift at a distance of 5 m from the sprayer. They found a correlation between pesticide deposition and wind, temperature, driving speed, and vapour pressure deficit. Our experiments show that deposition in trials employing low-drift nozzles (TeeJet) are generally lower than that in trials using flat fan nozzles. The AI nozzle is registered as a 50–75% drift-reducing nozzle in Germany. In Weisser’s experiments,23) the reduction was even larger. The AI nozzle in general produces larger droplets than the Hardi nozzle. Therefore, a large part of the spray liquid is expected to sediment, and less pesticide is available for spray drift. This is underlined by Nuyttens et al. (2007)19,29) and many others1,11,12,30) who have studied the effect of equipment and equipment setting, primarily on horizontal deposition. As an example, it has repeatedly been shown that different nozzle types give rise to different depositions, which is in line with our results for the two nozzle types employed in the present study.
It is known that spray drift is closely related to the droplet size distribution of the spray,31) and this is sustained by the significant influence of initial droplet size on spray drift found in this project. Several factors may affect droplet size, including nozzle type, nozzle pressure and temperature.19) Therefore, it seems reasonable to suggest that a more general description of spray drift can be based on the formation and behavior of different droplet size classes under varying environmental conditions. Droplet size characteristics can be obtained under controlled conditions in wind tunnels or in environmentally controlled chambers, as done in this project and described by Nuyttens et al. (2009).32) However, such measurements do not take the effects of turbulence and other weather conditions into account. This factor is implicitly included in experiments such as the one described in the present project, which is one of the reasons field experiments cannot be entirely substituted with laboratory experiments.
Generally, spray drift experiments have a high variance between samples from the same positions, as described in Arvidsson.28) However, the fact that the statistical model established in this study describes 82% of the variance among samples from five field trials suggests that many of the most important factors have been included in the model and that it may be worthwhile to try to account for the remaining variation.
In the calculation of the deposition in the hedgerow, we used the accumulated deposition from three spray swaths (36 m); however, under normal conditions this is not sufficient, since fields are usually much wider. For instance, in Denmark the average field size is approximately 4 ha.33) Therefore, it is important to be able to predict the contribution from more spray swaths in order to make trustworthy risk assessments, and the present study has shown that at least five spray swaths (60 m) add significantly to the total deposition, especially at the top part of the hedgerow.
Several aspects concerning spray drift have not been covered by the present project but should be included in future estimates of spray drift, e.g. other types of pesticides, additives, other types of spraying equipment (air-assisted sprayers in particular), and hedgerow characteristics such as species composition and porosity.
The authors thank laboratory technicians without whom this study could not have been done. The work of this paper was funded by the Danish Environmental Protection Agency.
