Characterizing RPAS coverage at four cardinal points on a vertical plane: Practical implications for spraying wheat at T3

Context

Global research into Remotely Piloted Aerial Systems (RPAS) is producing pesticide residue, drift and efficacy data that is helping to inform federal regulatory policy. It is reasonable to assume that Canada will ultimately sanction the use of RPAS for agricultural spraying. The first registered products will likely be fungicides intended for broad acre crops such as soybean, corn, and wheat.

Those considering RPAS for agricultural spraying have expressed interest ranging from general curiosity to high demand. Successful adoption will be contingent on expectation management, which in turn requires education on the functional differences between RPAS and conventional application technologies.

Quadrotor RPAS design dominates the current commercial landscape, with typical models featuring four rotary atomizers and 40 L tanks. There have been improvements in recent years, but these designs continue to suffer from a low rate of productivity (by North American standards for broad acre crops). This is due, in part, to low volumetric capacity and limitations with rotary atomizer design, which result in a debatably short effective swath width. Broadly, “swath width” refers to the minimal span consistently sprayed by a single pass, while “effective” indicates a spray coverage (i.e. deposition pattern and threshold dose) sufficient to achieve the desired result.

Our research efforts have focused on identifying and evaluating variables that influence effective swath width. These include operational settings such as altitude, travel speed, volume applied and nozzle settings. They also include environmental factors such as meteorological conditions, crop morphology, and planting architecture. Establishing a combination of settings that account for these factors will inform operator practices and optimize the balance between RPAS effectiveness and efficiency.

Study Objective

Fusarium head blight is a significant economic threat in wheat. Fungicide application takes place at the T3 stage of development, with the intention of providing panoramic coverage of the wheat head. RPAS is being considered to apply these fungicides.

The pursuit of productivity tempts operators to push operational settings to the point that spray coverage is compromised. This study will use operational settings based on the results of previous work and assign flight speed as the independent variable.

Coverage will be assessed using water sensitive paper (WSP) positioned at the top of the canopy and oriented vertically in four cardinal directions to emulate the circumference of the wheat head. Isolating the resultant coverage in each cardinal direction may provide insight into droplet behaviour within the RPAS spray cloud and perhaps better assess effective swath width.

Coverage from a conventional field sprayer will also be characterized. This represents the current standard and it will provide a basis for comparison.

Materials and Methods

Site

The experiment was conducted at 45939 John Wise Line, St Thomas, Ontario (42.7320746, -81.0879887) on June 1, 2025. Common seed wheat was planted on October 6^th, 2024, at 1.8 million seeds/ac on a 19 cm (7.5 in) row spacing. At the time of spraying, wheat was at the T3 stage of development, approximately 0.7 m (2.5 ft) high.

Holder Design

3D-printed holders were designed in Autodesk Fusion. They feature tabs that create a pressure fit (for quick WSP loading and unloading) and a back support (to prevent WSP movement in a downwash) to vertically position 1×3” WSP facing out on four cardinal points (figure 1). 21 poles were positioned on a 1 m spacing, leaving the tops coplanar with the wheat heads for in situ swathing. The square cross-section of the poles corresponded to a square depression in the WSP holder, ensuring the samplers were correctly aligned (figure 2). The gcode file to 3D print your own holders can be downloaded here. We used PLA filament with 3 walls and a 15% rectilinear infill to print 16 holders at a time on a Bambu P1S.

Wind direction is indicated by the direction the wind is coming from, not the direction it’s blowing to. Therefore, a wind blowing in a northern direction is referred to as south wind. The WSP holders were aligned east to west, perpendicular to the prevailing southern wind. We established a trampled path ~0.75 m on the downwind side. This left the wheat canopy surrounding the samplers intact while still permitting access to the WSP holders (figure 3).

WSP were given unique serial numbers to identify their position, and pre-loaded into the holders. 170 loaded holders were stored on raised grids inside four shallow plastic bins, arranged in the order they would be used (figure 4). This greatly expedited placement and retrieval, allowing for more repetitions to take place while optimal weather conditions held.

Application Method

A DJI T50 RPAS maintained and calibrated by the cooperator was fitted with four rotary atomizers. Previous work has demonstrated that higher rates of flow can have a detrimental effect on the spray quality from rotary atomizers, so distributing the flow over four nozzles was intended to prevent this. The atomizers applied a 250 µm spray over a 7 m swath width, as selected on the flight controller. Altitude was 3 m above the wheat heads, and flight speed was either 6 or 10 m/s (three repetitions each). Application volume was 50 L/ha, anticipating this to be a future label requirement.

The RPAS flew a racetrack pattern over the samplers (figure 5). It flew with a prevailing tailwind between samplers 14 and 15, and then back with a prevailing headwind between samplers 7 and 8. We employed a DJI RTK-1 base station, which claims 1 cm horizontal accuracy.

The RPAS was given sufficient distance (~40 m before and after the samplers) to reach the target speed, which was confirmed with a screenshot from the flight controller. The RPAS tank (capacity 40 L) contained municipal water with 0.125% v/v of NIS (Ag-Surf II). The level of liquid in the RPAS tank was maintained at 40 L throughout the trial to eliminate the effect of a changing payload.

The field sprayer, maintained and calibrated by the cooperator, was a New Holland SP 275s. It extended the left boom over the samplers from position 1 through to 12 and made a single pass with a prevailing tailwind (three repetitions). It traveled at 4.5 m/s (16 km/h or 10 mph) spraying 187 L/ha (20 gpa) and rough terrain caused the boom height to fluctuate between ~25 cm and ~50 cm above the wheat heads. The nozzles were Greenleaf Technologies TADF 06 (greys) on a 50 cm spacing operated at 50 psi to produce a Coarse spray quality. The tank contained municipal water with 0.125% v/v of NIS (Ag-Surf II).

Trial Procedure

WSP holders were placed just prior to spraying while the RPAS or field sprayer was positioned ~40 m beyond the samplers. When wind conditions were deemed appropriate, a signal was given to initiate spraying. On pass completion, one minute elapsed before initiating collection to permit complete deposition of the spray and drying of the droplets.

Weather Data

Weather data was collected using a Kestrel 3550AG weather meter (Kestrel Instruments) in a vane mount positioned roughly 1 m below RPAS altitude. Data was logged as the RPAS or field sprayer boom passed the samplers (table 1). In the case of the RPAS, there was very little difference between the two passes per repetition, so values were averaged.

Digitization

WSP (Spot-On) were scanned using a DropScope^TM (SprayX). The software reported droplet density and percent area coverage, but only deposit density is considered in this report.

Results

Comprehensive Observations

When the average deposit density from each of the four WSP per holder is added, we have a measure of total panoramic coverage. The mean total panoramic coverage from three repetitions of each treatment is shown in figure 6. For the RPAS treatments, distinct coverage peaks typical of RPAS applications correspond to the flight passes through poles 14-15 and 7-8. There is a slight spray displacement due to an occasional shift to a west wind (i.e. overall coverage shifted towards pole 1). However, the entire spray swath appears to have fallen within the range of the samplers. The racetrack flight pattern produced higher coverage at the interface between passes (pole 10) compared the extremes (beyond poles 19 and 3) suggesting some overlap. Deposit density was higher for the slower RPAS flight speed, and RPAS produced a higher deposit density than the field sprayer. Field sprayer coverage data is included for perspective but is generally not referred to unless it has some bearing on the evaluation of the RPAS coverage.

Swath Width

If the coverage at pole 10 represents the edge of each swath, then a swath width of ~6.0 m can be estimated based on similar coverage at poles 16-17 and 3-4. This is less than the programed value of 7.0 m.

This inference is supported when these averaged values were entered in an Excel-based model that calculates swath width. The model sums deposits from adjacent swaths assuming a racetrack pattern. Threshold coverage is subjective but adhering to the objective of establishing a balance between over- and under-dosing with the lowest possible C.V., we calculated swaths between 5.0 and 6.5 m (table 2).

It was expected that the field sprayer would produce a somewhat trapezoidal coverage pattern, tapering up at pole 12 (boom extreme) and level to pole 1 (sprayer chassis). Instead, note the gradual increase in coverage from pole 12 to 1. This is likely the result of a swaying boom, where the boom end swayed more than the point closest to the sprayer. If this degree of coverage represents the industry standard, it is notable that the average boom coverage is either on par with, or considerably less, than that of the RPAS.

Deposit Density

The RPAS was programmed to produce a 250 µm droplet size, while the field sprayer produced Coarse (~218-349 µm) droplets. Smaller stains were produced by the RPAS than the field sprayer (figures 7 and 8), and their circular/oval shape suggest both a smaller droplet volume and a somewhat perpendicular flight path. Droplets produced by the field sprayer left long streaks, which suggest higher droplet volumes and a more parallel flight path.

When the overall average coverage is calculated the RPAS at 6 m/s deposited an average 700 drops/cm² and an average 400 drops/cm² at 10 m/s. The field sprayer deposited an average 280 drops/cm². When volume remains constant, smaller droplet diameters produce a greater number of droplets than with larger droplet diameters, so more droplets would be expected from the RPAS. However, the RPAS applied only 50 L/ha while the field sprayer applied 182 L/ha. Therefore, the RPAS distributed a greater density of potentially higher-concentration droplets on each WSP holder compared to the field sprayer. Further, a slower flight speed deposited a higher density of stains than a faster flight speed.

Effect of Travel Speed

The 6 m/s treatment resulted in slightly smaller swath widths (1 m or 15% less) than the 10 m/s treatment. The positive relationship between swath width and flight speed has already been established. The positive relationship between flight speed and off target drift has also been established, which may account for the significantly fewer deposits (almost 50% fewer) in the swath at 10 m/s versus 6 m/s. When a higher deposit density is valued, such as in the case of contact fungicide application, the loss of productivity from a slightly smaller swath width is a reasonable compromise for the superior coverage within that swath.

Coverage by Cardinal Point

North Facing WSP

This side of the WSP holder faced north, away from the prevailing south wind. The RPAS deposited far more on this face when traveling with the wind between poles 14-15, making this side face the RPAS retreat (figure 9). It suggests the RPAS blows down and back, even against a tail wind. This is supported when RPAS returned between poles 7-8 and deposited comparatively less on this face.

RPAS speed may have had an effect. With airblast sprayers, slower travel speeds produce greater dwell times, which increases the distance a droplet travels on a given trajectory. Logically, when flying away from the retreat face, higher speeds would impart a greater forward momentum on a droplet, cancelling out some of the backward momentum (watch a video here). If this were the case, there would be comparatively improved deposit density on the 14-15 pass for slower speeds and reduced coverage between poles 7-8 as it blew past the target. Figure 9 supports this hypothesis.

South Facing WSP

This side of the WSP holder faced south, into the prevailing south wind. Considering the 6 m/s treatment, we see more coverage on the 7-8 pass than the 14-15, because it represents the retreat side of the sprayer (figure 10). The differential is far less on this plane than the north facing (figure 9) because the prevailing wind likely blew spray into the WSP on the 14-15 pass. Nevertheless there is significantly more on 7-8.

This relationship is less clear for the faster 10 m/s treatment. There does tend to be higher deposit on the 14-15 pass as spray was blown into the collectors. However far more was expected on this face for the 7-8 pass as it represents both the retreat face and has the added benefit of wind. Further, there was far less coverage overall when compared to the slower flight speed. We have no explanation for the lack of coverage on the 7-8 pass and can only conclude that higher speeds left droplets airborne and were not conducive to coverage.

West Facing WSP

This side of the WSP holder faced west, into the slight west wind. We see coverage is almost exclusively on those WSP facing the drone (figure 11). In other words, as the RPAS passed between 14-15 and 7-8, coverage was positively skewed from this point. The skewed coverage was evident at both flight speeds, but overall coverage was higher for the slower speed. Once again, we cannot explain why there was significantly reduced coverage on the 10 m/s pass between poles 7-8 except to suggest the spray may have remained airborne.

East Facing WSP

This side of the WSP holder faced east, away from a slight west wind. We see that coverage is almost exclusively on those WSP facing the drone in the upwind direction (figure 12). Overall coverage was slightly higher for the slower speed, but far less overall coverage compared to the west facing samplers (figure 11). This is likely because the light west wind caused spray to displace from pole 21 to pole 1, washing past the back (insensitive) sides of the WSP.

Summary

RPAS can provide par or better panoramic wheat head coverage compared to a conventional ground rig when they are flown using reasonable operational settings in optimal environmental conditions. A moderate flight speed (~6-8 m/s), appropriate altitude (~3 m above wheat heads), and four rotary atomizers producing a Medium-Coarse (~250 um) droplet size can produce an in situ 6 m swath width at 50 L/ha. Higher flight speeds produce a marginally wider swath at the cost of reduced droplet density and increased drift potential.

Assessing coverage using vertical WSP facing four cardinal points has provided insight into the behaviour of spray from an RPAS. Droplets from any application technology tend to deposit with wind and gravity, but rotor downwash represents an additional variable unique to RPAS. That force, combined with a forward cant of the drone during flight, lead droplets to deposit on vertical surfaces that face the rear (retreat) of the sprayer as well as surfaces that face and intercept spray that radiates laterally from the flight path.

Acknowledgements

Drone Spray Canada, Bayer Canada, volunteers Kurtis Pilkington and Natalie, and grower-cooperator Adam Pfeffer are gratefully acknowledged for their contributions to this study.