|
Spray Drift, Operator Exposure, Crop Residue and Efficacy: Early Indications for Equivalency of Uncrewed Aerial Spray Systems with Conventional Application Techniques
Jane A. S. Bonds1,*, Naresh Pai2, Sarah Hovinga2, Katie Stump3, Rebecca Haynie4, Sheila Flack2, Travis Bui5
Published in Journal of the ASABE 67(1): 27-41 (doi: 10.13031/ja.15646). Copyright 2024 American Society of Agricultural and Biological Engineers.
1Bonds Consulting Group, Panama City, Florida, USA.
2Bayer Crop Science, Chesterfield, Missouri, USA.
3CropLife America, Washington, District of Columbia, USA.
4Syngenta International AG, Basel, Basel-Stadt, Switzerland.
5Corteva Agriscience, Indianapolis, IN, USA.
*Correspondence: jasbonds@gmail.com
The authors have paid for open access for this article. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License https://creative commons.org/licenses/by-nc-nd/4.0/
Submitted for review on 24 April 2023 as manuscript number MS 15646; approved for publication as a Review Article by Associate Editor Prof. Emilio Gil and Community Editor Dr. Heping Zhu of the Machinery Systems Community of ASABE on 5 September 2023.
Highlights
Abstract. Uncrewed Aerial Spray Systems (UASS) are being adopted at a rapid pace in agricultural applications of crop protection products. The data required to effectively regulate their use must be gathered to position UASS in terms of equivalency with other conventional practices. In Fall 2021, the CropLife America Drones Working Group initiated an effort to collect published information on establishing the equivalency of UASS applications with conventional application types as it relates to spray drift, operator exposure, crop residue, and efficacy. Based on the published literature, our comparison demonstrated that UASS spray drift is lower than aerial, higher than ground boom, and similar to orchard airblast applications. However, this comparison is based on limited data and needs further confirmation. Individual use cases and other application variables will need to be considered to determine if this generalization applies (e.g., adjuvant use, rotor and nozzle configuration, etc.). For operator exposure, this work supported the current consensus that applications with UASS have less potential for exposure in some respects (e.g., compared to backpack applications), but for other job steps that are unique to UASS (such as mixing and loading) more information is needed. With respect to crop residue, UASS applicators follow the label for conventional application techniques with the same directions for use (i.e., application rate, pre-harvest interval, and number of applications), but there is no evidence that pesticide residues resulting from a UASS application are any different to conventional application techniques. In terms of efficacy, applications with UASS tend to be equivalent to conventional methods; however, more information is needed, especially where good coverage is a requirement. The assessment of published literature on UASS demonstrates potential equivalency in certain key areas and supports the responsible use of this emerging technology, while more information on spray distribution within the target zone, off-target droplet movement, operator and bystander exposure, and pesticidal efficacy continues to be generated.
Adoption of Uncrewed Aerial Systems (UAS) into the civil sector and their regulation has proceeded at a rapid pace (FAA, 2019). Uncrewed Aircraft Spray Systems (UASS) have been estimated as a significant sector of the overall UAS market in multiple economic evaluations (Sachs, 2016). Application of crop protection products using UASS provides several benefits, including applications in undesirable topography or soil conditions, spraying small irregular-shaped fields, or when airplanes and helicopters are too expensive to use. While the use of UASS continues to grow around the world, especially in Southeast Asia, the number, size, and distribution of operators continue to remain small within the United States. A combination of a lack of commercially viable use cases and regulatory hurdles is likely among the factors limiting the growth of this market.
From a regulatory standpoint, one of the areas of exploration is the comparison of off-target movement through spray drift of UASS applications relative to conventional spray platforms (e.g., ground boom, aerial, and airblast applications). For conventional platforms, the characteristics of spray drift (e.g., role of droplet size) and corresponding sensitive parameters (e.g., wind speed) are relatively well understood from several decades of field testing (Rautmann et al., 1995; Salyani and Cromwell, 1992) and the development of predictive models (Bilanin et al., 1989; Ellis and Miller, 2010). It is generally assumed that the spray drift from a UASS would be within the range of the drift curves for conventional aerial and ground-based applications. While this appears to be a reasonable approach, supporting it with field data will help to substantiate this assumption and build consensus.
Previous knowledge of the safety and performance of conventional applications can be leveraged along with summaries of recent data generated for UASS, for instance, to understand the off-target movement characteristics or to inform an interim drift model specific to UASS. Other areas in need of consensus in terms of UASS equivalency to conventional platforms include exposure to operators and bystanders, as well as residues in crops and efficacy. These are important considerations for regulatory approvals of crop protection products but haven’t yet been summarized based on previous literature.
Due to the nascent stage of UASS technology, published studies exhibit a diversity in data collection and reporting methodologies that are often specific to achieving study objectives. Several workstreams are underway to improve this situation (e.g., OECD, 2021), but there are currently no standardized protocols for testing UASS performance for gaining regulatory approvals. One complicating factor is that while spray systems in conventional platforms are relatively stable in terms of overall design, UASS platforms continue to evolve rapidly to better meet the needs of the market (Sachs, 2016). This evolution is certainly needed to grow the UASS market but presents hurdles for regulators as they contend with the potential human and ecological risks associated with the delivery of crop protection products using these new systems.
The solution to this issue could be two-pronged. First, defining a benchmark test UASS system—one that encompasses various designs (e.g., number of rotors), application methods (e.g., nozzle types), and capacity (e.g., payload)—could help establish consensus within the scientific and regulatory community. Such a benchmark system, when tested, would conservatively account for the risks associated with other UASS platforms. Second, a consistent approach to UASS field data collection is needed to allow for its broader use and comparison across platform types, spray characteristics, and environmental conditions.
Defining a benchmark platform or developing a standardized testing protocol is outside of the scope of this article. However, important steps have been taken in this article to support these endeavors through: (1) Gathering published studies that measured spray dispersal of crop protection products using UASS, especially those that simultaneously also compared them to conventional application types; (2) Summarizing information related to spray drift, crop residues, operator, bystander exposure, and efficacy when compared to that of conventional application types; (3) Identifying key platform characteristics and application parameters investigated to date to provide a delineation for a benchmark sprayer; and (4) Providing guidance on the key parameters that should be recorded in UASS studies to allow for better utilization in regulatory risk characterization.
Ultimately, the goal of this article is to summarize the state of scientific and published literature on spray drift, human (bystander and operator) exposure, crop residues, and efficacy as it relates to UASS applications while providing context on how they compare to conventional applications.
Methods
Peer-reviewed scientific literature, relevant internal reports, and web-based resources were searched to identify research related to pesticide application with UASS. Topic search terms, used individually and in combination, included ‘Drone-Unmanned-Uncrewed-Remotely-Piloted-Aerial Spray Systems’, ‘Drift’ ‘Control’ ‘Efficacy’, ‘Residues’, ‘Off Target’, ‘Insecticide’, ‘Herbicide’, ‘Fungicide’, ‘Dispersal’ ‘Operator’, ‘Bystander’ and ‘Exposure’. The Web of Science, Google Scholar, Google, and Mendeley databases were all used to explore the field, as were citations within identified papers. Searches took place during Dec 2021-Jan 2022.
The literature is diverse. It reflects the spectrum of study approaches that range from computer simulations to single observation field trials to extended multiyear field studies. Small-scale field studies can provide a baseline from which to identify and further characterize how UASS compares to conventional methods. Extrapolations or inferences on the relevance and reliability of some of the studies are presented with caution. In this area, there is a lack of agreed-upon guidelines or testing protocols to standardize the conduct of these trials. There were frequently experimental aspects that limited the robustness or reliability of the product to being useful in a regulatory context. Therefore, only 33 of the 97 studies pertaining to UASS obtained for this review were considered both relevant and reliable for regulatory purposes. The primary focus of this review was on spray drift and gathering raw data to prepare an empirical database for the development of interim curves for UASS compared to conventional techniques.
Spray Drift
For many application types, major efforts are directed at the development of relevant spray drift models. Empirical models, for example, are typically employed for the prediction of off-target movement for ground application equipment, and the beginnings of such a data collection are presented here.
There are a number of studies with UASS that have followed the ISO standard for the measurement of spray drift (ISO, 2005). This standard establishes principles for the measurement of droplet drift from many equipment types designed for the application of pesticide products. For example, the deposition onto horizontal surfaces at different distances downwind is relevant to the risk assessment of contamination of surface water or exposure to non-target plants or animals. Whereas the measurement of airborne profiles is relevant to risk assessments relating to inhalation effects and bystander exposure. Where data are limited, as with UASS, comparative assessments of the relative drift profiles from different application systems are needed, and studies that incorporate conventional reference spraying systems are the most valuable.
These ISO-compliant studies provide data as a “percentage of applied,” which is useful for normalizing between applications with different application rates. The data currently available can provide some information on the overall position of UASS compared to other application types, but it highlights the need for more detailed test protocols to be developed. One of the limitations of ISO 22866 (2005) is that it suggests that samplers collect drift down to a representative distance where 90% of the spray has been collected, which may not be the preferred method in some cases. It is understood that regulators within the US, for example, would prefer to see a near-analytical limit of detection numbers as opposed to the ISO-recommended 90%. In addition, many published studies with UASS do not measure spray drift further than 50 m from the field edge, which is less than previous studies that are used in current regulatory datasets, which tend to extend beyond 100 m.
An early drift study by Xue et al. (2014) placed deposition samplers (Mylar cards) at the furthest distance downwind as compared to other studies (100 m). The samplers were set at distances of 2, 4, 6, 8, 10, 20, 50, and 100 m downwind. The aircraft operated at a height of 5 m, a flight speed of 3 m/s, and at a wind speed of 3 m/s. Deposition drift accounted for 12.9% of the total spray volume, while 90% of the drift was concentrated within the first 8 m downwind of the sprayed area. These results and the guidance of ISO 22866 are likely the reasoning for the subsequent more nearfield distances.
As with all application equipment, spray drift studies involving UASS are conducted with cross winds that reposition the swath downwind. With crewed aircraft, the swath off setting is well-understood, and applicators either adjust the last flightline upwind half a boom width or turn off the downwind half of the boom (Thomson et al., 2013). The same offsetting is likely required for UASS, yet it has not been incorporated into any operating procedures. A system for calculating swath width and location after wind driven offsetting is described in Wang et al. (2020a). They observed that almost none of the centers of spray deposition for a single swath aligned with the corresponding flight path, meaning that the spray deposition from the UASS sprayer of all treatments deposited downwind in different degrees under the effect of the natural crosswind. The authors noted that this effect was more marked with the Fine nozzle compared to the Coarse, low-drift, air induction nozzle. As with all application equipment, droplet size is the primary driver for controlling drift. Table 1 is a reference for the droplet size categories used throughout this document.
| Table 1. Droplet size classification based on ISO 25358. | |
| Size Classification | Dv0.5[a] Range (microns) |
| Extremely Fine | < 60 |
| Very Fine | 61-105 |
| Fine | 106-235 |
| Medium | 236-340 |
| Coarse | 341-403 |
| Very Coarse | 404-502 |
| Extremely Coarse | 503-665 |
| Ultra Coarse | > 665 |
[a]DV0.5 refers to the midpoint droplet diameter (median), where half of the volume of spray is in droplets with a smaller diameter, and the other half of the volume is in droplets with a larger diameter than the median. | |
Wang et al. (2020c) compared the drift potential of three different fine droplet size distributions (Dv0.5) of 100, 150, and 200 µm with centrifugal nozzles repeated over a range of meteorological conditions using a four-rotor (XAGP20) UASS with a 4 m flight height and a 5 m/s flight speed. Samplers extended to 50 m downwind, where deposition amounts were lower than the detection limits of 0.0002 µL/cm2. The distance downwind containing 90% of the applied spray was on average 16.5, 13.2, and 12.1 m for 100, 150, and 200 µm, respectively. Based on the results of this study, the drift distance of this specific UASS configuration is described by the author as less than that of manned aerial applications. Chen et al. (2020) also investigated the effect of droplet size on in-swath distribution and downwind drift. The nozzles were flat fan nozzles 110 01, 110 015, 110 02, and 110 03 (with DV0.5 95, 121, 147, and 185 µm) with a cumulative drift of 74, 50, 36, and 23% of applied, respectively, showing that, as with conventional equipment, droplet size is still the primary driver of drift.
As with conventional spray equipment, it has also been shown that incorporation of appropriate adjuvants can reduce drift. A study with three different single rotor UASS (Wang et al., 2018b) showed that adjuvants with anti-drift properties did in fact reduce spray drift. Compared to water, Silwet DRS-60, ASFA+B, T1602, Break-thru Vibrant, QF-LY, and Tmax could reduce spray drift by 65%, 62%, 59%, 46%, 42%, and 19%, respectively. Wang et al. (2021) incorporated the rotors of a UASS to provide downwash in a wind tunnel study measuring droplet size distributions for a range of commonly used nozzles. The effects of adjuvants were also investigated over a range of temperature and humidity. The results showed that adjuvants can be effectively used for drift reduction especially under high-temperature and low-humidity conditions.
In the US, Brown and Giles (2018) conducted a study in vineyards with the large single-rotor RMAX® and a fine nozzle. The deposition averaged 0.4% of the applied amount at 7.5 m downwind and 0.03% at 48 m downwind; approximately 82% of the drifted material was estimated to deposit within the first 7.5 m downwind of the field edge. The drift deposition curves from this study indicate a lower fraction of downwind drift compared to conventional fixed-wing or rotary-wing aircraft (Salyani and Cromwell, 1992), as well as certain ground sprayer configurations in orchards and vineyards (Fox et al., 1998; Grella et al., 2017).
As with crewed aerial applications, the flight height and windspeed play an important role in managing downwind drift. Studies with the larger single-rotor UASS (3WQF120-12; Anyang Quanfeng Biological Technology Co., Ltd.) applied a medium droplet size, a forward speed of 3 m/s, and operated at heights of 1.5, 2.5, and 3.5 m (Wang et al., 2018a). Each treatment height was run with a higher and a lower wind speed. At each treatment height, the higher wind speed led to less mass in-swath and increased drift. When the UASS height was lower than 2.5 m and the mean speed was less than 2.82 m/s, the authors reported that 90% of the total measured spray drift was within 10 m (Wang et al., 2018a).
The effect of nozzle position is also important, as with other application equipment. With crewed aerial applications, the nozzles are typically positioned within 75% or 80% of the wing or propeller diameter with fixed and rotary aircraft, with many labels now specifying 65%. With orchard airblast applications, specific nozzle placements or baffles are advised to reduce droplet sizes and trajectories that would be susceptible to drift. The same is true of the UASS nozzle position; studies by Bonds et al. (2022) showed that placing nozzles outside of the propeller diameters led to a reduction in swath uniformity and an increase in drift. Richardson, Rolando et al. (2020) conducted studies on the impact of nozzle position on improved deposition characteristics. The study highlighted the sensitivity of nozzle position in relation to flight direction. The DJI (SZ DJI Technology Co., Ltd., Shenzhen, China) Agras MG-1 drone used typically positions the nozzles directly under the four side rotors. Moving the nozzles to the front or back showed less deposition when the nozzles were positioned on the leading edge versus the trailing edge. Another study by Wang et al. (2018b) looked at three single-rotor aircraft, showing a similar increase in drift when the nozzle distance from the boom centerline was the rotor radius. In this study, the three UASS were operated at 1.5-2 m height and at 4-5 m/s. All sprayers were operating with fine spray nozzles; the primary difference between the sprayers was the length of the boom. The boom lengths were described as percentages of the rotor diameter, which were 98%, 58%, and 56%. The drift, described as percentages leaving the target zone, were 24%, 9.4%, and 2.4% of the total spray volume, respectively. This requires further investigation, but it is possible that off-target losses will decrease if the spray is released within 75% of the rotor diameter, as with manned rotary aircraft.
Herbst et al. (2020) investigated the effect of both droplet size (fine and coarse) and sprayer type: a single rotor, 6 and 8 rotor UASS with nozzles on a boom inside the prop diameter, and a second 8 rotor with four nozzles under the rotors. Initially, the authors concluded that, in the arable system, the UASS sprayer type had little impact on spray drift. However, the DJI Agras MG-1 model with the nozzles positioned directly under the four side rotors, as opposed to on a boom positioned within the rotor diameter, did show a small increase in potential drift with monofilament lines at near the edge of field (2 m from the spray edge) at a 1.5-m altitude (surrogate boom sprayer model). This was followed by a marked increase in drift with the DJI model and the fine spray quality at a 3.5 m altitude (surrogate orchard sprayer model). Table 2 provides a summary of the experiments discussed above, showing the drift distances to where 90% of the spray is accounted for.
| Table 2. Summary of UASS spray drift trials that took a measure of distance to recovery of 90% of applied. | ||||||
| Author | Altitude (m) | Travel Speed (m/s) | Droplet Size Classification | Rotor | Wind Speed (m/s) | Distance to 90% (m) |
| Wang, 2018a | 1.5 | 3 | Medium | 1 | 0.7 | 6.9 |
| 1.5 | 3 | Medium | 1 | 2.2 | 3.9 | |
| 2.5 | 3 | Medium | 1 | 1.8 | 3.7 | |
| 2.5 | 3 | Medium | 1 | 4.7 | 10.1 | |
| 3.5 | 3 | Medium | 1 | 1.8 | 33.5 | |
| 3.5 | 3 | Medium | 1 | 3.7 | 46.5 | |
| Xue et al., 2014 | 5 | 3 | Medium | 3 | 8 | |
| Brown and Giles 2018[a] | 4.6 | 5.4 | Fine | 1 | 3 | 7.5 |
| Wang, 2020c | 4 | 5 | Very Fine | 4 | 2.8 | 16.5 |
| 4 | 5 | Fine | 4 | 2.9 | 13.2 | |
| 4 | 5 | Fine | 4 | 2.0 | 12.1 | |
| Wang, 2018b | 2 | 5 | Medium | 1 | 6.3 | 10 |
| 2 | 5 | Medium | 1 | 7.3 | 6.4 | |
[a]The distance in this study was 82% of applied, not 90%. | ||||||
Interim Drift Curves
In a comparative assessment of drift data from UASS and conventional applications, Herbst et al. (2020) investigated four different UASS, all operating at a speed of 2 m/s with nozzles delivering coarse and fine droplet size distributions. Two UASS heights were tested: a height of 1.5 m represented a bare ground arable model (ground boom), and a height of 3.5 m above the ground with a 2 m artificial canopy represented a vineyard model (airblast sprayer). Drift samplers were Petri dishes positioned at 3, 5, 10, 15, and 20 m downwind. For the arable model, system drift from the coarse nozzle was equal to the standard drift curve (Van de Zande et al., 2015) for boom sprayers, while the fine nozzle was higher. For the vineyard model system, drift from the coarse nozzle was lower than the standard drift curve, while drift from the fine nozzle was comparable to the standard curve for vineyards (Van de Zande et al., 2015).
Similar results were observed by Anken and Dubuis (2020) when drift was measured from two UASS (HSE Agrofly, DJI Agras) operating at a height of 3-3.5 m over a range of wind speeds with a mixture of nozzles delivering fine and coarse droplet size distributions. Drift samplers were measured out to 50 m and compared to standard drift curves (Rautmann et al., 2001). In this study, both the fine and coarse nozzles on the UASS were shown to have lower drift than the corresponding curves for orchard airblasts. Compared to the standard curve for tractor boom sprayers, drift from the coarse nozzles on the UASS was lower and from the fine nozzle was higher.
These complementary drift studies offer insight into the relative drift volumes and distances following spray application with UASS. Data indicates that for UASS, sampling beyond 50 m would not be a useful expenditure of resources, but high-resolution sampling within swath and down to the first 20 m downwind is critical to better understanding the details around swath offsetting and drift. In recent years, the number of platforms with the Real Time Kinematics (RTK) Global Positioning System (GPS) has been increasing, thus reducing the likelihood and magnitude of misapplication and enabling researchers to focus on wind shift. With crewed aircraft, this shift is well understood, and applicators know to offset a distance upwind defined by the models to ensure swath placement within the target zone. Similar data and education materials will be necessary for uncrewed applications.
In an effort to begin to organize the available information, an empirical database for UASS drift was created (fig. 1). This database includes only raw, original data, not averaged data. Data sets were provided by A. Herbst (Germany) and J. Bonds (US); more relevant data may be available, but raw data was not provided.
 |
| Figure 1. Flow chart representing the data required for inclusion into the empirical database. |
Figure 1 provides a breakdown of the data required from a spray drift trial. The data categories include trial information, trial conduct, equipment data, and replicate data. Note that there are numerous data points related to the application rate and the appropriate calibration of the aircraft. It is essential that such application data be collected because there are numerous instances in the literature where a lack of this data makes the results unreliable. Within the equipment section, the intended application rate should be noted along with the nozzle type, size, number, and pressure, along with the assigned swath width. The Replicate Data section requires information on how much was discharged, the spray time, and the distance the spray was on. The in-swath deposition profile is needed to estimate the swath width.
From this database, drift data are presented as an average of the percentage applied, measured at each of the downwind distances. The first set of data from Germany collected in collaboration with China Agricultural University is presented in figures 2 and 3. This data was collected in 2018 and 2019. The aim of the studies was to assess the impact of UASS platform design and droplet size distribution for drift in a steep-sloped vineyard. The platform designs were a single rotor, a six rotor and two eight rotors. Three of the platforms had nozzles on a boom, while one of the eight rotor platforms had the nozzle positioned under the four side rotors. The droplet size distributions were fine and coarse. These studies led to the additional work conducted in Germany, which supported the decision to permit the use of UASS in Germany, as further explained in the Discussion section (Herbst et al., 2020). The first set of studies (2018) were conducted at a low altitude (1.5 m), which was compared to a boom sprayer paradigm. The second set of studies (2019) were conducted at a higher altitude (3.5 m) over an artificial vineyard canopy, which was compared to an airblast paradigm. To provide context, the European Union (EU) basic drift curves for boom (fig. 2) and airblast (fig. 3) sprayers are presented to begin to position UASS against conventional equipment (Rautmann et al., 2001). Figure 2 shows that, compared to a boom sprayer, both the fine and coarse nozzles produced more drift than the EU standard curve. Figure 3 shows that, compared to the airblast sprayer, the fine nozzle drifted slightly more and the coarse nozzles slightly less than the EU standard curve.
Raw data was also accessed from a set of studies conducted in 2021. The aim of this data collection was to assess swath width, displacement, and drift for the application of larvicides in mosquito control (Bonds et al., 2022). Three nozzle types were used throughout the 2020 field trials, including a reference nozzle, the TeeJet XR 110 03 flat fan, a Turbo TeeJet Induction flat fan (TTI) 110 01, and a Lechler IDK 110 01 flat fan. The platform was a PV35 (Leading Edge New Smyrna, FL) 6-rotor UASS with a payload capacity of 15 kg, with the four nozzles positioned on a boom -1.4 m -0.4 m +0.4 m +1.4 m from the center line. The UASS manufacturer recommended swath width for applications made at an altitude of 4.5 m was 5 m for the PV35 with a flight speed of 3.6 m/s and the total system flow rate of 425 ml min-1 required to achieve an application rate of approximately 28 L/ha.
  | ||||||||
| (a) | (b) | |||||||
| Figure 2. Drift assessment with four different UASS applied at 2 m/s travel speed and 1.5 m altitude compared to the EU basic drift curve for ground boom sprayers applied with two different drop size: (a) fine drop size, and (b) coarse drop size. | ||||||||
 | ||||||||
| (a) | (b) | |||||||
| Figure 3. Drift assessment with four different UASS applying at 2 m/s travel speed and 3.5 m altitude compared to the EU basic drift curve for orchard airblast sprayers applied with two different drop sizes. (a) Fine drop size, and (b) coarse drop size. | ||||||||
The Flat Fan 110 03 (FF) nozzle produced a medium spray quality, and the air induction low drift nozzles (IDK and TTI) produced a coarse spray quality (table 1). These data have then been presented with the same EU standard drift curves used to show that this UASS drift data aligns with the work conducted by Herbst et al. (2020) for boom and orchard airblast sprayers. In addition, the basic drift curve for aerial, used by the Australian Pesticides and Veterinary Medicines Authority (APVMA, 2017) was added to cover the three primary application types (fig. 4).
This data and this literature review, however, are primarily for use within a US regulatory context, so the data have also been presented (fig. 5) with three standard tier I curves extracted from AgDRIFT (Teske et al., 2002): the regulatory model used by the Environmental Protection Agency (EPA). The curve for boom sprayers is for the 50th percentile, low boom, and fine to medium/coarse droplet size distribution. The orchard airblast spray drift curve within AgDRIFT is a composite of multiple orchard types but does allow for variation in droplet size distribution. The standard curve for aerial use is a fine-to-medium droplet size distribution.
The last comparison of UASS drift curves to regulatory data sets is the drift curves utilized by Health Canada’s Pest
Management Regulatory Agency (PMRA) which is responsible for pesticide regulation in Canada. The Canadian regulators utilize empirical curves for tractor boom sprayers generated by Wolf and Caldwell (2001) using a Medium nozzle. The curves for orchard airblast were generated in Germany by and the data for orchard early stage was chosen. The aerial curve is generated by AGDISP 8.21 using the Canadian default scenario for field crops using a Medium nozzle (fig. 6).
With existing methods of application, the data currently in the database effectively positions spray drift from UASS between crewed aircraft and ground boom sprayers. With the EU Basic Drift Curves, the German data with fine and coarse nozzles, and the US data with medium and coarse nozzles, UASS spray drift sits in line with the EU standard drift curve for orchard airblast (figs. 4 and 5). The US data, when compared to the AgDRIFT curves, shows drift volumes close to orchard airblasts in the near field, then bisects ground at 10-15 m downwind and is less than ground boom sprayers at 20 m (fig. 6). Lastly, the data compared to the PMRA standard curves shows that the UASS drift curves above ground boom are clearly less than the aerial and orchard airblast curves.
 |
| Figure 4. UASS drift assessment and comparison with the APVMA Basic Drift curve for aerial and the EU Basic Drift curves for Orchard Airblast and Ground Boom. The six-rotor UASS operated at a velocity of 3.6 m/s and 4.6 m altitude, with medium and two coarse nozzles. |
 |
| Figure 5. UASS drift assessment and comparison with the EPA AgDRIFT 2.1.1 Orchard, Ground, and Aerial Curves. The six-rotor UASS operated at a velocity of 3.6 m/s and 4.6 m altitude, with medium and two coarse nozzles. |
 |
| Figure 6. UASS drift assessment and comparison with the PMRA AgDisp for aerial, and empirical data for tractor boom sprayers with a medium spray distribution and orchard airblast early. The six-rotor UASS operated at a velocity of 3.6 m/s and 4.6 m altitude, with medium and two coarse nozzles. |
Bystander Exposure
Bystander exposure has been assessed in several studies that incorporated monofilament lines, which measure the airborne fraction of dispersed pesticides. The monofilament lines are erected at various heights from the ground and distances from the edge of the field. The spray volume collected by monofilaments, however, tends to be artificially high because they are not corrected for sampling rate. The correction for sampling rate is complex, requiring droplet size, wind speed, and temporal data, which are rarely collected. This means that the numbers reported in these studies should only be used as a comparative measure between treatments within a particular study, rather than to compare between studies.
Wang et al. (2020a) investigated the effect of a fine (TR 80 067) and coarse (IDK 120 015) droplet size on potential spray drift collected on monofilament line samplers at 2 m from the edge of field with three different UASS included: a single-rotor (3WQF120-12), a six-rotor (3WM6E-10), and an eight-rotor (3WM8A-20) aircraft flown at 2 m/s and 3.5 m height above the crop (simulated vineyard 1.8m). At the lowest height on the monofilament lines (0.5 m), the highest airborne deposition was obtained with the fine spray in the order of eight-rotor (142% of applied), followed by the single-rotor (121%), and the six-rotor (84%). The coarse spray produced significantly less potential drift: the percentage leaving the target zone was 14% with the single-rotor, 13% with the eight-rotor, and 6% with the six-rotor UASS. Herbst et al. (2020) integrated the downwind sedimentation drift and the potential drift on monofilament lines at 2 m from the edge of the field. In general, the airborne spray drift in vineyard applications was higher than in the arable crop scenario; this difference was due to release height (3.5 m versus 1.5 m, respectively). The hollow cone nozzles (fine) versus the air induction nozzles (coarse) released significantly more spray from the target area, increasing the chance of bystander exposure.
Xue et al. (2014) measured airborne drift data from monofilament lines strung at heights between 0.5 and 4 m at distances of 2 and 50 m downwind. For the monofilament lines placed at 2 m, the lowest lines collected the highest volumes of the descending spray cloud; the 0.5 m height was 14.6%, and at 4 m height was 4.8% of applied. At 50 m monofilament distance, the detected amount was almost zero. In a separate study, Wang et al. (2020c) utilized monofilament lines at 2 m and 12 m from the edge of the field every 1 m up to a 5 m height to investigate the effects of a Dv0.5 of 100, 150, and 200 µm. The platform was a quadcopter (P20, XAG) operating at a relatively high altitude of 4 m and a forward speed of 5 m/s. The airborne drift on the monofilament lines for each treatment generally increased as the line sampling height decreased. At the 2 m distance, the 100 µm droplet size at wind speeds above 3 m/s resulted in deposits between 40% and 60% of applied; with winds below 3 m/s deposits of 20% of applied were detected. As droplet size increased (150 and 200 µm) and wind decreased, so did deposition on the lines. All depositions at 12 m were less than 20% of applied at the 1 m sampler height with the 100 µm droplet size and less than 10% with the 150 and 200 µm droplet size distributions.
Wang et al. (2018a) conducted a drift study in a pineapple crop using a single-rotor (3WQF120-12) UASS operated at a fixed velocity of 3 m/s at 1.5, 2.5, and 3.5 m heights above the canopy with a medium droplet size distribution of 268 µm, repeated over a range of wind speeds. Monofilament lines were positioned at 10, 25, and 50 m from the edge of field, with lines at the heights of 5, 2, and 1 m. At the low UASS operating height (1.5 m) and under low wind speeds (0.5-2.2 m/s), deposition measured on monofilament lines was close to zero. At the medium flight height (2.5 m), measurable deposition (0.01 µg/cm2) was observed at 10 m from the edge of the field at the higher wind speed. At the 3.5 m UASS operating height, the wind speed varied from 1.0 to 5.1 m/s, and deposition was low but measurable at all distances (0.005-0.03 µg cm2).
Some may question the adequacy of monofilament lines for drift sampling, suggesting that high-resolution air samplers are superior. Brown and Giles (2018) utilized high resolution air samplers in a drift study but found both air samples (15 and 48 m) were below the method detection limit.
The studies tend to agree that at 2 m from the edge of the field, between 15% and >100% of applied can be captured on monofilament lines, defining what is being lost. As was mentioned in the drift section, >100% of the applied is likely due to a wind-driven offset of the actual swath past the edge of the field. At 12 m overall, less than 20% of the applied material was collected, and at 50 m, deposits were near zero. These data should help shape the design of future experiments in terms of the useful placement of samplers for bystander exposure and the correct positioning of the flight line to keep the last swath within the edge of field.
Operator Exposure
To better understand the risks of exposure to operators and field staff, information is needed on the potential for exposure to residues on equipment and from tasks such as operating, mixing, loading, maintaining, cleaning, and transport. Operator exposure will depend on factors like their physical location compared to the sprayer, crop type, equipment, training, and atmospheric conditions. The exposure of UASS operators is assumed to be less than that of operators with a knapsack sprayer because the operators are separated from both the crop (target area) and the sprayer. Exposure of a UASS operator is more likely than that of an operator sealed within the cab of an aircraft or tractor. Comparative exposure to operators without a cab is unknown. A simple understanding of the distance from, the duration of the application, and the protections in place for the applicator could make the position of UASS operator exposure a relatively simple academic exercise. In terms of task differences, Borysenko, Antonenko et al. (2021) showed that exposure potential was higher for the UASS refueler compared to the pilot. This difference was due to longer exposure times for refuelers, with an exposure time of 42-45 min versus 28-30 min for the pilot. Moreover, the pilot is separated from direct contact with the spray, which reduces risk to a level no more than that of an observer, about 2-3 orders of magnitude less than the refueler (Fargnoli et al., 2019). Compared to the other application techniques, most UASS and knapsack sprayers need to be refilled by hand, whereas the tractor boom, airblast, and crewed aircraft have closed transfer systems.
In the case of residues on the sprayer, qualitative observations and numerical simulations show the spray to have multiple trajectories, including an upward component that could accumulate active ingredients on the aircraft (Zheng et al. 2018). An accumulation of residues could also be incurred by flying back through a fine spray cloud that has yet to settle, meaning the equipment itself could present an increased risk. Other concerns revolve around the potential for increased risk of sensitization or irritation due to higher concentrations of active ingredients with ULV and LV sprays and should be considered for UASS operators and field staff.
Li et al. (2020) measured deposition on the UASS during an almond orchard application by placing filter papers on the sprayer boom, body, and arms. Swabs from the rotor blades were also collected. Recovery numbers were considered low (< 6 µg), but the exposed area was not reported. It was noted, however, that the spray boom and drone arms had the highest residues, and since the drone arms are used by the operators to lift the aircraft, wearing proper personal protection equipment (PPE) is important. Following an application in an apple orchard, Liu et al. (2020) measured the residues of the active ingredient on the surfaces of both a UASS and an airblast sprayer. The residues were measured on filter paper located on the fan casing (air blast) or battery (UASS), the front and back of the tank, and on the tractor or airfoil. The average residue on the UASS was 13.84 µg/cm2, compared to 0.58 µg/cm2 on the airblast sprayer. The airblast sprayer operated at 1058 L/ha, while the UASS operated at 60 L/ha. This potentially reflects the higher concentration of the pesticide solution in the UASS but cannot explain the difference in surface residues alone. The residues were 23 times higher on the UASS compared to the airblast, whereas the dilution rate was only 17 times less. The increase may be due to an increase in deposition to surfaces due to previously noted turbulence as well as the increase in concentration. Regardless of how deposition occurred, if there is an increase in the concentration of pesticide residue, this could have consequences for personal protection during cleaning and maintenance.
Yan et al. (2021) conducted a whole-body operator exposure measure for UASS and knapsack applications in cowpea. In terms of crop structure, cowpea is a tall (2 m) thin wall of vegetation. The UASS flew over the top of the canopy while the knapsack operators swung the lance from top to bottom while walking into the spray. The UASS operator was 2 m upwind of the edge of the field, applying 4 L over a time frame of 1.3 min. The knapsack applicators were applying 15 L over a time frame of approximately 12 mins. This scenario represents a near-worst-case exposure for a knapsack sprayer operator. The knapsack operators were exposed to 1.95 g/kg, compared to 0.35 g/kg for the UASS operator.
These studies show that UASS can accumulate larger quantities of active ingredient on the vehicle compared to conventional techniques, likely due to the concentration of the spray and the turbulent airflow increasing contact with the spray. In comparison to knapsack sprayers, however, the separation in space and the reduction in exposure time will potentially lead to lower exposure. For conventional crewed aircraft and ground sprayers, there is little reporting on comparative differences in exposure to the operator. For the mixers and loaders, the conventional crewed aircraft and ground sprayers have engineering controls (e.g., closed transfer) in place to further protect both the applicator and the field staff that are not commonly available for UASS.
Crop Residues
Currently, UASS applicators follow the label language for conventional equipment so that the volume of active ingredient applied, the number of applications, the preharvest intervals, etc. are equal. The only potential deviation is the volume of carrier, and, therefore, the concentration of active(s) in the spray is increased (Martinez-Guanter et al., 2020). These studies initially compare low-volume UASS coverage to conventional application techniques, followed by residue analysis. Monitoring the residue persistence is crucial to assessing if the concentrated UASS sprays (lower carrier volumes) dissipate at rates similar to conventional, more dilute sprays. This evaluation helps determine their impact on pre-harvest intervals and maximum residue limits (MRLs).
Meng et al. (2018) showed no difference in the initial residue or half-life (2 hours to 14 days) of imidacloprid for control of wheat aphid between UASS applications (12.6 L/ha) and knapsack applications (270 L/ha), indicating that application rates do not affect the initial volume or half-life of the compound. Similarly, Li et al. (2020) conducted applications with chlorantraniliprole (111 g a.i./ha) for the control of navel orange worm, comparing UASS applications (46.8 L/ha and 93.6 L/ha) to an orchard airblast sprayer (935 L/ha) in almonds. When the same product use rate was applied at 111 g a.i./ha (0.099 lb a.i./ac) in the drone and ground applications, the resulting active ingredient concentrations by the drone applications at 46.8 L/ha and 93.5 L/ha were respectively 10 and 20 times higher than that of the air-blast ground application applied at 935 L/ha (Li et al., 2020). The almonds were sampled for pesticide residues alongside filter papers and water-sensitive papers to characterize the spray distribution in the canopy. The percentage of coverage was greater with the high volume of the airblast sprayer at 12% compared to the 93.5 L/ha (4%) and the 46.8 L/ha (2%) application rates, but overall pesticide residue levels on whole, un-hulled almonds were equal. Further studies in alfalfa showed similar levels and distribution of residues observed with both crewed and uncrewed aerial application methods, indicating that crop residue tolerances (MRL’s) will not be impacted (Li et al., 2021).
There is not much information on the pesticides and their breakdown products with UASS applications in the literature. There may be, however, a potential increase in residue information as countries set up protocols for Good Agricultural Practice (GAP) and the associated increase in residue testing requirements (Matthews, 2019). The current literature does show that where the rate of active ingredient is equal, the concentration per target structure is also typically equal. Therefore, regardless of coverage and other deposition characteristics, where the label is followed, residues should be comparable to conventional methods.
Efficacy
There were several efficacy studies identified throughout this review. Many studies compared the control of a specific pest by UASS to accepted norms for efficacy. Others conducted investigations into the physical characterization and the effect of different application settings to improve coverage. For this review, however, only studies that made comparisons to a conventional application were considered. These studies focused primarily on investigating if the lower volumes applied with UASS produced comparative control and if the downwash or the addition of adjuvants improved coverage.
Comparison to an Industry Standard Knapsack
There are several studies available that compare the low volume UASS application to high volume knapsack sprayers. Generally, where the target pest was an insect, control with a UASS was likely to be comparable to conventional techniques as pest mobility reduces the need for good coverage. When the target is a fungal pathogen, good coverage is essential, especially with contact fungicides. The lower application rates with UASS were less effective than higher rates and conventional techniques.
Hu et al. (2021) compared a four-rotor UASS (3WQFTX-10) with application rates of 30, 22, and 18 L/ha and application heights of 1, 1.5, and 2 m to a conventional knapsack sprayer at 450 L/ha for cotton aphid control. The bioassay data showed that one day post treatment, the knapsack sprayer returned significantly higher mortality. On day three, all treatments were equal between the knapsack and the UASS at the two upper application rates and the two lower altitudes. On day seven, the two lower altitudes and the knapsack sprayer were still equal, with the higher altitude returning significantly lower efficacy data. Overall, aphid control positively correlated with volume applied and droplet density on the underside of the leaves, which is the preferred aphid habitat. Yan et al. (2021) used the biopesticide spinetoram to control thrips in cowpea, a 2 m tall vine crop. The treatment volumes were 600 L/ha for the knapsack compared to 22.5, 30, and 37.5 L/ha with the UASS. Thrip control on the first day was 69% (knapsack) and 70%, 80%, and 81% for increasing UASS volumes, respectively. The UASS (DJI T16) was more effective than applications with a knapsack (3WBD-18). Additionally, the authors noted that the UASS applications greatly reduced water consumption and working time, which is attractive for laborious knapsack applications (Yan et al., 2021).
Meng et al. (2018) investigated the effect of low carrier volume (13 L/ha) with a single-rotor UASS (3WQF120-12) to a high-volume knapsack sprayer (260 L/ha). The authors were also investigating the effect of dose reduction (imidacloprid at 90 g a.i./ha and 72 g a.i./ha) with two different adjuvants (organosilicon or methylated vegetable oil) to improve coverage with the low volume UASS applications. The first study investigated preventative control and showed that the dose could be reduced with the UASS application without a loss of control with the addition of organosilicon adjuvant (82% control) as compared to the standard knapsack treatment (87% control). A subsequent study on aphid-infested crops showed that after 14 days, there was no difference in aphid control between the full dose and the reduced dose with organosilicon applied by UASS and the knapsack control (91, 90, and 92% control, respectively); all these treatments had significantly higher control (P < 0.05) than the reduced dose treatment without adjuvant and the reduced dose with the methylated vegetable oil (87% and 89%, respectively).
Qin et al. (2018) investigated the effect of an ULV application for mildew control in wheat. The UASS utilized an application rate of 15 L/ha, while the knapsack sprayer applied at a rate of 300 L/ha. This translated to UASS applied dose rates of 270, 360, and 450 g/ha (triadimefon SC 44%) and 450 g/ha for the knapsack sprayer. Ten days after application, control with the UASS was lower than the knapsack sprayer: 68% for the UASS at the highest dose and 73% for the knapsack (P < 0.05). Good coverage is essential for fungal control, and the authors suggested the addition of an adjuvant to improve coverage and retention of the compound.
Wang et al. (2020b) compared a four-rotor (TAX) UASS to an electric knapsack sprayer (450 L/ha) application, investigating the effects of UASS spray volume (9 and 18 L/ha) and tank-mix adjuvants for the control of rice blast and leaf roller. The UASS operated at 2.0 m above the crop (at the panicle initiation stage), with the volume adjusted via flight speed (6 and 3 m/s). Increasing the spray volume and adding an adjuvant (methylated crop oil) significantly (P < 0.01) increased droplet density, percentage coverage, and control of rice blast and rice leaf roller for the UASS application. Among all treatments, the UASS at 18 L/ha with an adjuvant returned the best rice blast control efficacy of 63%. For the rice leaf roller, control efficacy was high, ranging from 84% to 96% for the UASS at 18 L/ha, which was not significantly different from the knapsack sprayer at >96%. Overall, this suggests the control of fungal pathogens is more challenging with ULV applications but can be overcome with the addition of an appropriate adjuvant.
The type of active ingredient and volume applied can also have an impact. Optimal control efficacy of wheat aphid and powdery mildew was achieved at >16.8 L/ha volume with systemic insecticide, and at 28.1 L/ha with contact insecticide and fungicide. The control achieved with the UASS treatments was equal to the conventional knapsack sprayer. For the knapsack sprayer, the high spray application rate of 450 l/ha led to run-off, and a lower spray volume of 225 L/ha achieved better deposition and control efficacy for both contact and systemic insecticides (Wang et al., 2019).
Comparison to an Industry Standard Ground Applications
Wang et al. (2019) compared a six-rotor UASS using an application rate of 10 L/ha (3WTXC8-5) to three standard application methods (boom sprayer at 300 L/ha, knapsack at 300 L/ha, and mist blower at 75 L/ha), measuring both the spray distribution and biological efficiency against wheat aphids (imidacloprid at 86 g a.i./ha). The spray coverage area differed among treatments, but the deposition of active ingredients was comparable across all sprayers tested. Coverage with the UASS was significantly lower (2%) compared to the tractor boom, mist blower, and knapsack, which achieved 38%, 17%, and 21%, respectively. The UASS also had reduced canopy penetration compared to the higher-volume applications, which led to the lowest losses to the ground. The UASS deposited 0.13 µg/cm-2 to the soil surface compared to the boom sprayer at 0.39 µg/cm-2. Aphid control was positively correlated with carrier volume. The UASS had significantly lower control than the other treatments on days 1 and 7, but was equivalent on day 3. On day 7, control with the UASS was 71%. The authors still considered this acceptable, especially when the relative working efficiency of the application methods was considered. The working efficiency of the UASS was 4.1 ha/hr, the boom sprayer 2.4 ha/hr, the mist blower 1.6 ha/hr, and the knapsack 0.2 ha/hr.
Lou et al. (2018) compared a UASS spraying 12 L/ha with a conventional boom spraying 450 L/ha, for the control of aphids and spider mites. Five days after treatment, the level of control observed was 90% (boom sprayer) and 64% (UASS) for aphids, and 68% (boom sprayer) and 61% (UASS) for spider mites. Also, Xiao et al. (2020) investigated four different application methods for the control of Fusarium Head Blight (FHB) and mycotoxin contamination in wheat. The UASS and mist sprayer improved the control efficiency 14.2%-19.6% compared with the traditional knapsack and boom sprayer. A dose-dependent response effect was also observed, but the increases were not statistically significant. Comparisons of the mist sprayer and UASS revealed no statistically significant differences in terms of FHB control, but the working efficiency of the UASS was 13.3 times greater than that of the mist sprayer.
UASS studies conducted by Liao et al. (2019) investigated defoliation in cotton, which can be a dense and challenging canopy. The authors compared three UASS (YR-GSF06 with four rotors, TXA with six rotors, and YR-AU 15 with eight rotors) at three different application rates to a tractor boom sprayer applying 180 L/ha. The UASS application rates were changed with pressures of 200, 300, and 400 kPa and corresponded to respective application volume rates of 48, 72, and 96 L/ha. Application volume was the main treatment parameter, returning roughly 2%, 5%, and 10% coverage, respectively. Although there were clear differences in terms of percent coverage with changes in application volume, all UASS applications achieved high levels of defoliation and were higher than the tractor boom sprayer. The application volumes used by the UASS were within the recommended range of 48 L/ha for manned aerial and 96 L/ha for ground boom applications; however, the 180 L/ha application rate used by the tractor boom sprayer may have been too high.
Li et al. (2020) conducted their studies in large, dense almond tree canopies. The UASS applied a carrier volume of 46.8 L/ha and 93.6 L/ha, compared to an orchard airblast sprayer applying a carrier volume of 935.4 L/ha (chlorantraniliprole 111 g a.i./ha, plus Dyne-Amic surfactant 0.06% v/v). The large Yamaha RMAX model spray release height was maintained between 1.8-2.4 m with a flight speed of 1.3 m/s. The percentage of coverage was greater with the high volume of the airblast sprayer at 12% compared to the 93.5 L/ha (4%) and the 46.8 L/ha (2%) UASS application rates. There were distinct differences in residue patterns at different canopy elevations between the aerial and ground application methods, with the UASS depositing more to the upper canopy and the airblast sprayer to the lower canopy. No difference in control was seen between treatments, mainly because damage was low; this meant it was not possible to statistically separate treatments. There were additional studies in orchards that showed lower coverage but retention of the same rate of active ingredient when compared to an industry standard (Liu et al., 2020; Tang et al., 2018).
These studies demonstrate that, in terms of efficacy, the UASS appears to be comparable to the larger ground rigs, with most studies being conducted in orchards. Within orchards, UASS could prove advantageous, especially where pests are distributed at the tops of the trees. One of the advantages of UASS is its ease of use and improved working efficacy, especially in comparison with a knapsack sprayer. This is less clear with the high-capacity ground rigs. Martinez-Guanter et al. (2020) conducted an economic assessment of a UASS compared to an orchard airblast in olive groves. The conventional equipment cost and working rate were 26 ?/h and 4.8 ha/h, compared to 12.3 ?/h and 5.4 ha/h for the UASS. This shows that UASS can be considered as effective and economical as large ground rigs; however, it is recommended that the enthusiasm toward these new delivery systems be measured.
Comparison to an Industry Standard, Aerial
The only field study that makes a direct comparison between crewed and uncrewed aircraft was conducted by Li et al. (2021) in alfalfa. Their work in field trials with chlorantraniliprole showed insect control, with a significantly reduced number of lepidopteran pest species at all sampling dates regardless of aerial application method (UASS vs. crewed aircraft) compared to the untreated control (P < 0.05) and providing excellent larval control. Larval pest control was at least 90% or greater compared to the untreated control, and the efficacy and residue results provided additional confidence that UASS applicators may follow current agrochemical company label recommendations for manned aerial application and achieve adequate control of alfalfa pests when following the label GAPs.
Applications with UASS and crewed aircraft both benefit from the ability to apply compounds in a timely fashion and can utilize their downwash to potentially improve coverage of the target surface. The primary difference between the two is logistical capacity. A traditional fixed-wing airplane can carry between 1500 and 3000 L of spray liquid and travel across large fields at a wide range of speeds between approximately 40 and 80 m/s, while multi-rotor UASS sprayers, normally equipped with a smaller spray tank (10-40 L), travel at a much lower speed (2-8 m/s). Without sufficient field data and comparative studies producing a baseline reference between UASS and crewed airplane applications, it is difficult to analyze the gaps and strengths of the uncrewed aerial application technology compared to that of crewed applications (Li et al., 2021). As more real-world examples accumulate, backed by rationale for use by the farmer (e.g., economics, ease of use, user safety, etc.), the distinction and benefit of these two technology types will become clearer.
Physical Characterization
Due to the many and varied systems available, optimization of UASS technology would be aided by the development of a user-accessible, mechanistic model describing spray droplet trajectories following their release from multi-rotor UASs under different environmental conditions and application settings. The literature to date shows that unique challenges in terms of the basic mechanics of spray dispersion from UASS are unlikely, as the physics are the same. Both crewed and uncrewed aerial vehicles produce a strong downwash, which pushes the spray quickly toward the ground. This may potentially provide better distribution over individual plants as opposed to merely coating their upper surfaces (Teske et al., 2018). If the height and speed for any given platform is too low, it could create a ground effect which will force the downwash up (Bonds et al., 2022). These conditions are well understood for crewed aircraft but not for UASS systems. This lack of understanding of downwash bounce is exacerbated by the wide range of UASS configurations (e.g., number of rotors, size of rotors, and rotor position in relation to nozzle location) and potential operating conditions (e.g., flying speed and release height) (Richardson et al., 2019). Unique to crewed aerial applications is the development of the Mechanistic Agricultural Dispersal Model (AGDISP), which enables a more sophisticated prediction of off-target movement. This model is used to conduct regulatory risk assessments related to spray drift from manned aerial applications and the subsequent demarcation of no-spray buffer zones for certain compounds. It remains to be seen whether this existing exposure model will work for uncrewed systems due to their small size, slow speeds, and varying styles. For example, the AGDISP model is specific for single-rotor helicopters and fixed-wing planes and is not yet validated for the multiple rotors seen with many UASS. It is possible that the large single-propeller and fixed-wing UASS, when modeled, may work, but validation from empirical data would still be required.
The multiple rotors present a unique challenge with respect to increased turbulence because the trajectories of the droplets and the variation around those trajectories are complex. There have been numerous attempts to model UASS using Computational Fluid Dynamics (CFD), underscoring the complexity of the turbulent wake (Guo et al., 2020). A few of these studies have incorporated both spray and forward motion but no functional model development (Zhang et al., 2020). Some steps have been made towards a mechanistic model – the Comprehensive Hierarchical Aeromechanics Rotorcraft Model (CHARM), which attempts to simulate the UASS. The complexity of the UASS wake is especially critical to understand. One of the primary conclusions was that as flight speed increases, a critical speed (hover downwash speed) is reached at which the downwash transitions to outwash, and this can occur well before the released droplets reach the ground, potentially increasing drift (Teske et al., 2018). The Charm+AgDISP solution and the multiple CFD models have made some progress in this space, but further development and validation will be needed which in turn will depend on robust and consistent field data collection. This manuscript provides data from previous studies to help identify pit falls, and information gaps, and it also lays out the parameters that need to be recorded in future studies.
Towards a Benchmark Platform
The purpose of a benchmark platform is to define a hypothetical UASS that, when field tested, would give the regulators confidence that it would conservatively cover the remaining platforms from a safety (environmental and human) and performance (efficacy) perspective. Within this data-gathering exercise on equivalency, a database of the various application parameters used in the literature was created (table 3). The continued development of such a database could help define a benchmark platform for the development of testing and comparative testing protocols.
An electronic version of the database is posted on this website: [https://bit.ly/ASABE_MS15646]. The database contains UASS from 38 different manufacturers. The database records the author and year of each study, along with a brief description of the experimental protocol and a notation as to whether the study was relevant and reliable or relevant but not reliable. The model number and the number of rotors for each experimental platform are provided. Single-rotor and 8-rotor platforms tend to be the larger, high-capacity systems, which is where the general trend in design is going. The general specifications of the platform are also provided, describing the dimensions, net weight, gross weight, and payload capacity.
| Table 3. The drone parameter database stores information from the literature on spray system parameters and experimental treatments. | |||
| Parameter | Min | Max | Mode |
| Rotor number | 1 | 8 | 1 |
| Payload L | 20 | 24 | 10 |
| Altitude m | 1 | 6 | 3 |
| Velocity m/s | < 2 | 9 | 3 |
| Swath Width m | 0.5 | 7 | 5 |
The experimental parameters frequently reported were spray altitude and velocity. Since the precision of the data were variable, they were aggregated into 1 m increments. The measured and/or assumed swath widths were also many and varied, ranging from 0.5 to 7 m. The boom length, boom location, nozzle spacing, and number are not consistently reported in UASS literature. This lack of consistency poses challenges for the successful identification of a benchmark, as with manned aircraft, the nozzle positioning will be crucial for determining deposition uniformity and drift. Approximately half of the platforms positioned the nozzles on the boom, and the other half attached under the rotors. The application rates in UASS studies ranged between 8-151 L/ha. The upper limit of this application rate was achieved with a slow forward speed of 1 m/s, which is not realistic for commercial applications. Other treatment settings and measures, such as flow rate and pressure, are recorded, along with a few measures of hover downwash from simulation studies.
Discussion
Regulators, pesticide companies, and applicators require confidence that UASS can deliver safe and effective applications that are at least comparable to existing commercial equipment and contribute to the environmental and economic goals of society and customers. Increased market adoption is expected to occur as tank capacity and battery longevity improve, and, most importantly, growers are assured that UASS spray technology will be economical and reliable in their business (Li et al., 2021).
There is a crucial need for focused research on characterizing spray distribution within the target area, understanding off-target droplet movement, assessing operator and bystander exposure, and evaluating pesticide efficacy under different pest pressures. Further field evaluations and real-world feedback are required to evaluate drone performance and to show that they are at least comparable to existing commercial equipment. One detail that was underscored in the recent OECD State of the Knowledge Literature review (OECD, 2021) is the need for training to ensure that experiments are conducted appropriately. There are several studies that are not designed properly because the relative importance of input parameters is not well understood. Data from factorial orthogonal designs fall prey to this in most instances, especially where the study is not appropriately replicated. In addition, drift studies should employ good laboratory practices whenever possible in field data collection and analytical methods to minimize the uncertainty resulting from collecting drift at low rates. This highlights the need for developing consistent experimental procedures to be authored and disseminated, which are actively being worked on by key stakeholders.
In terms of equivalency for spray drift, three independent research studies determined that spray drift from UASS is more than ground boom sprayers and less than crewed aircraft, aligning closely to the basic drift curves for orchard airblast for both the EU and the US. The parameters and physics that drive drift for conventional applications are the same with UASS, meaning that the placement with orchard airblast is appropriate as the droplets are released into ambient air above the crop with highly variable trajectories and velocities. Based on these data, German authorities have permitted fungicide applications by UASS in steep-sloped vineyards (Glaser et al., 2021; JKI, 2021). These are exclusively for products that have already been authorized or permitted for use with crewed helicopters in steep-sloped vineyards. Moreover, approved application parameters are limited to those nozzles and platforms held within their empirical database. That is, the UASS may not operate more than 2 m above the crop, and the flying speed must not exceed 3.6 m/s. The application can only be carried out with drones that can fly automatically (i.e., the route specified by the user, the speed, the height above the crop, and the switch-on and switch-off positions during spraying all must be automatic). There is a list of approved UASS that must operate with coarse (injector) nozzles and only for the application of pesticides in steep-sloped vineyards (JKI, 2021). Using a pesticide with UASS requires the additional authorization of the concerned authorities and must observe the regulations of aviation law as with crewed aircraft.
The literature underscores a need for a better understanding of bystander exposure, though equivalency studies do not indicate a potential increased risk. Operator exposure has been shown to be significantly less than from conventional knapsack applications. While there are a few studies that compare UASS to other conventional techniques, more information is needed regarding physical control measures, such as closed vs. open transfer systems. More information is needed on safety procedures and personal protection during cleaning, maintenance, and transport. In terms of residues, researchers in the published literature have followed the pesticide label, and no study has shown an increase in overall residues if the label is followed. In addition, the degradation rate does not appear to be affected by concentration. In terms of general efficacy, UASS has been shown to be greater than or equal to conventional application techniques. Where data show that efficacy is less than from conventional methods, an increase in worker efficiency often leads the author to conclude that UASS is equivalent. This assessment and comparison of published literature of UASS demonstrates potential equivalency in certain key areas and supports the responsible use of this emerging technology, while more information gathering on spray distribution within the target zone, off-target droplet movement, operator and bystander exposure, and pesticidal efficacy continues to be generated.
In closing, it is pertinent to introduce the task force that has been developed to provide this information in a timely and collaborative fashion. In the previously mentioned OECD State of the Knowledge report, several recommendations were made along with the data provided in this report, including identifying regulatory knowledge gaps relevant for UASS pesticide application. In response to these recommendations, the pesticide registrant industry formed the Unmanned Aerial Pesticide Application System Task Force (UAPASTF) in the summer of 2022 to advance global regulatory acceptance and guidance of this technology. The goal of the UAPASTF is to generate information and data, where appropriate, for submission to regulatory authorities to support the use of pesticides with UASS for crop protection, public health, and other uses. Within this context, the UAPASTF focuses on informing estimates for off-site movement, determining potential operator/handler exposure, and assessing crop residue contribution to human dietary exposure in risk assessment and regulatory approval processes. Additionally, generated data will contribute toward the evaluation of existing regulatory models and/or the development of new UASS-focused models that estimate exposures in risk assessment and regulatory approval processes. The UAPASTF has developed a general off-site movement study protocol that has been reviewed by multiple regulatory authorities and plans to publicly share this protocol in an effort to improve the quality of off-site movement studies being done with UASS. Further, results and conclusions from the data generated by UAPASTF will be presented at technical and scientific conferences and then considered for submission for publication in a peer-reviewed journal at an appropriate time to help build confidence in its use, especially for environmental exposure modeling, and ensure accessibility to a larger audience. It will be important to follow these efforts to incorporate the learnings into future drone-specific off-site movement modeling activities.
Acknowledgments
This work was financially supported by CropLife America (CLA), and the Drones Working Group within the association contributed to its content. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement from the authors or their respective organizations.
References
Anken, T., & Dubuis, P. H. (2020). Technical characteristics and risks of drift during the application of phytosanitary products with drones in viticulture. Agroscope, 29th June.
APVMA. (2017). Spray drift risk assessment manual. Australian Pesticides and Veterinary Medicines Authority. Retrieved from https://apvma.gov.au/sites/default/files/publication/28076-28076-spray_drift_risk_assessment_manual_alt_text_final_6.pdf
Bilanin, A. J., Teske, M. E., Barry, J. W., & Ekblad, R. B. (1989). AGDISP: The aircraft spray dispersion model, code development and experimental validation. Trans. ASAE, 32(1), 327-334. doi:https://doi.org/10.13031/2013.31005
Bonds, J. A., Fritz, B. K., & Thistle, H. W. (2022). A field programme for the spray distribution of unmanned aerial spray systems (UASS) and the development of larvicide systems for vector control. Proc. Int. Advances in Pesticide Application, (pp. 105-115).
Brown, C. R., & Giles, D. K. (2018). Measurement of pesticide drift from unmanned aerial vehicle application to a vineyard. Trans. ASABE, 61(5), 1539-1546. doi:https://doi.org/10.13031/trans.12672
Chen, S., Lan, Y., Zhou, Z., Ouyang, F., Wang, G., Huang, X., . . . Cheng, S. (2020). Effect of droplet size parameters on droplet deposition and drift of aerial spraying by using plant protection UAV. Agronomy, 10(2), 195. doi:https://doi.org/10.3390/agronomy10020195
Ellis, M. C., & Miller, P. C. (2010). The Silsoe Spray Drift Model: A model of spray drift for the assessment of non-target exposures to pesticides. Biosyst. Eng., 107(3), 169-177. doi:https://doi.org/10.1016/j.biosystemseng.2010.09.003
FAA. (2019). FAA aerospace forecast: Fiscal years 2019-2039. Federal Aviation Administration. Retrieved from https://www.faa.gov/sites/faa.gov/files/data_research/aviation/aerospace_forecasts/FY2019-39_FAA_Aerospace_Forecast.pdf
Fox, R. D., Derksen, R. C., & Brazee, R. D. (1998). Air-blast/air-assisted application equipment and drift. Proc. North American Conf. on Pesticide Spray and Drift Management, (pp. 108-109).
Glaser, M., Ruppert, F., Becker, M., Stephan, F., Rainer, K., & Herbst, A. (2021). Results of the measurements for drift in the application with unmanned aerial vehicles (UAV). Deutsche Pflanzenschutztagung.467, p. 206. Julius Kühn-Institut.
Grella, M., Gallart, M., Marucco, P., Balsari, P., & Gil, E. (2017). Ground deposition and airborne spray drift assessment in vineyard and orchard: The influence of environmental variables and sprayer settings. Sustainability, 9(5), 728. doi:https://doi.org/10.3390/su9050728
Guo, Q., Zhu, Y., Tang, Y., Hou, C., He, Y., Zhuang, J., . . . Luo, S. (2020). CFD simulation and experimental verification of the spatial and temporal distributions of the downwash airflow of a quad-rotor agricultural UAV in hover. Comput. Electron. Agric., 172, 105343. doi:https://doi.org/10.1016/j.compag.2020.105343
Herbst, A., Bonds, J., Wang, Z., Zeng, A., He, X., & Goff, P. (2020). The influence of unmanned agricultural aircraft system design on spray drift. Journal für Kulturpflanzen, 72(1), 1–11. doi:https://doi.org/10.5073/JfK.2020.01.01
Hu, H., Ren, X., Ma, X., Li, H., Ma, Y., Wang, D., . . . Ma, Y. (2021). Control effect on cotton aphids of insecticides sprayed with unmanned aerial vehicles under different flight heights and spray volumes. Int. J. Precis. Agric. Aviat., 4(1), 44-51. doi:https://doi.org/10.33440/j.ijpaa.20200401.161
ISO. (2005). ISO 22866:2005: Equipment for crop protection - Methods for field measurement of spray drift. International Standards Organization. Retrieved from https://www.iso.org/standard/35161.html
Julius Kühn Institute. (2021). Permit of plant protection products for the application by drones in steep slope vineyards. Bundesamt für Verbraucherschutz und Lebensmittelsicherheit, Federal Office of Consumer Protection and Food safety. Retrieved from https://www.bvl.bund.de/SharedDocs/Fachmeldungen/04_pflanzenschutzmittel/EN/2021/2021-08-24_drones.html
Li, X., Giles, D. K., Andaloro, J. T., Long, R., Lang, E. B., Watson, L. J., & Qandah, I. (2021). Comparison of UAV and fixed-wing aerial application for alfalfa insect pest control: Evaluating efficacy, residues, and spray quality. Pest Manag. Sci., 77(11), 4980-4992. doi:https://doi.org/10.1002/ps.6540
Li, X., Giles, D. K., Niederholzer, F. J., Andaloro, J. T., Langa, E. B., & Watson, L. J. (2020). Evaluation of an unmanned aerial vehicle as a new method of pesticide application for almond crop protection. Pest Manag. Sci., 77(1). doi:https://doi.org/10.1002/ps.6052
Liao, J., Zang, Y., Luo, X. .., Zhou, Z. Y., Lan, Y. B., Zang, Y., . . . Hewitt, A. J. (2019). Optimization of variables for maximizing efficacy and efficiency in aerial spray application to cotton using unmanned aerial systems. Int. J. Agric. Biol. Eng., 12(2), 10-17. doi:https://doi.org/10.25165/j.ijabe.20191202.4288
Liu, Y., Li, L., Liu, Y., He, X., Song, J., Zeng, A., & Wang, Z. (2020). Assessment of spray deposition and losses in an apple orchard with an unmanned agricultural aircraft system in China. Trans. ASABE, 63(3), 619-627. doi:https://doi.org/10.13031/trans.13233.
Lou, Z., Xin, F., Han, X., Lan, Y., Duan, T., & Fu, W. (2018). Effect of unmanned aerial vehicle flight height on droplet distribution, drift and control of cotton aphids and spider mites. Agronomy, 8(9), 187. doi:https://doi.org/10.3390/agronomy8090187
Martinez-Guanter, J., Agüera, P., Agüera, J., & Pérez-Ruiz, M. (2020). Spray and economics assessment of a UAV-based ultra-low-volume application in olive and citrus orchards. Precis. Agric., 21(1), 226-243. doi:https://doi.org/10.1007/s11119-019-09665-7
Matthews, G. (2019). Pesticides: Residues in crops and their application. Outlooks Pest Manag., 30(2), 85-87. doi:https://doi.org/10.1564/v30_apr_10
Meng, Y. H., Lan, Y. B., Mei, G. Y., Guo, Y. W., Song, J. L., & Wang, Z. G. (2018). Effect of aerial spray adjuvant applying on the efficiency of small unmanned aerial vehicle for wheat aphid control. Int. J. Agric. Biol. Eng., 11(5), 46-53. doi:https://doi.org/10.25165/j.ijabe.20181105.4298
OECD. (2021). Report on the state of the knowledge – Literature review on unmanned aerial spray systems in agriculture. OECD Series on Pesticides(No. 105). Paris: OECD Publ. Retrieved from https://www.oecd.org/chemicalsafety/pesticides-biocides/literature-review-on-unmanned-aerial-spray-systems-in-agriculture.pdf
Qin, W. C., Xue, X. Y., Zhang, S. M., Gu, W., & Wang, B. K. (2018). Droplet deposition and efficiency of fungicides sprayed with small UAV against wheat powdery mildew. Int. J. Agric. Biol. Eng., 11(2), 27-32. doi:https://doi.org/10.25165/j.ijabe.20181102.3157
Rautmann, D., Ganzelmeier, H., Spangenberg, R., Streloke, M., Herrmann, M., Wenzelburger, H.-J., & Walter, H.-F. (1995). Studies on the spray drift of plant protection products (Vol. 305). Blackwell Wissenschafts-Verlag.
Rautmann, D., Streloke, M., & Winkler, R. (2001). New basic drift values in the authorization procedure for plant protection products. Workshop on Risk Assessment and Risk Mitigation Measures (WORMM), 13, pp. 3-141.
Richardson, B., Rolando, C. A., Kimberley, M. O., & Strand, T. M. (2019). Spray application efficiency from a multi-rotor unmanned aerial vehicle configured for aerial pesticide application. Trans. ASABE, 62(6), 1447-1453. doi:https://doi.org/10.13031/trans.13509
Sachs, G. (2016). Drones: Reporting for work. Goldman Sachs Research.
Salyani, M., & Cromwell, R. P. (1992). Spray drift from ground and aerial applications. Trans. ASAE, 35(4), 1113-1120. doi:https://doi.org/10.13031/2013.28708
Tang, Y., Hou, C. J., Luo, S. M., Lin, J. T., Yang, Z., & Huang, W. F. (2018). Effects of operation height and tree shape on droplet deposition in citrus trees using an unmanned aerial vehicle. Comput. Electron. Agric., 148, 1-7. doi:https://doi.org/10.1016/j.compag.2018.02.026
Teske, M. E., Bird, S. L., Esterly, D. M., Curbishley, T. B., Ray, S. L., & Perry, S. G. (2002). AgDrift®: A model for estimating near-field spray drift from aerial applications. Environ. Toxicol. Chem., 21(3), 659-671. doi:https://doi.org/10.1002/etc.5620210327
Teske, M. E., Wachspress, D. A., & Thistle, H. W. (2018). Prediction of aerial spray release from UAVs. Trans. ASABE, 61(3), 909-918. doi:https://doi.org/10.13031/trans.12701
Thomson, S. J., Womac, A. R., & Mulrooney, J. E. (2013). Reducing pesticide drift by considering propeller rotation effects from aerial application near buffer zones. Sustain. Agric. Res., 2(3). doi:https://doi.org/10.5539/sar.v2n3p41
Van de Zande, J., Rautmann, D., Holterman, H. J., & Huijsmans, J. F. (2015). Joined spray drift curves for boom sprayers in The Netherlands and Germany. Wageningen, Plant Research International. Retrieved from https://edepot.wur.nl/353554
Wang, C. L., Zeng, A. J., He, X. K., Song, J. L., Herbst, A., & Gao, W. (2020). Spray drift characteristics test of unmanned aerial vehicle spray unit under wind tunnel conditions. Int. J. Agric. Biol. Eng., 13(3), 13-21. doi:https://doi.org/10.25165/j.ijabe.20201303.5716
Wang, C., Herbst, A., Zeng, A., Wongsuk, S., Qiao, B., Qi, P., . . . He, X. (2021). Assessment of spray deposition, drift and mass balance from unmanned aerial vehicle sprayer using an artificial vineyard. Sci. Total Environ., 777. doi:https://doi.org/10.1016/j.scitotenv.2021.146181
Wang, G. B., Li, X., Andaloro, J., Chen, P. C., Song, C. C., Shan, C. F., . . . Lan, Y. (2020). Deposition and biological efficacy of UAV-based low-volume application in rice fields. Int. J. Precis. Agric. Aviat., 3(2), 65-72. Retrieved from https://doi.org/10.33440/j.ijpaa.20200302.86
Wang, G., Han, Y., Li, X., Andaloro, J., Chen, P., Hoffmann, W. C., . . . Lan, Y. (2020). Field evaluation of spray drift and environmental impact using an agricultural unmanned aerial vehicle (UAV) sprayer. Sci. Total Environ., 737, 139793. doi:https://doi.org/10.1016/j.scitotenv.2020.139793
Wang, G., Lan, Y., Yuan, H., Qi, H., Chen, P., Ouyang, F., & Han, Y. B. (2019). Comparison of spray deposition, control efficacy on wheat aphids and working efficiency in the wheat field of the unmanned aerial vehicle with boom sprayer and two conventional knapsack sprayers. Appl. Sci, 9(2), 218. doi:https://doi.org/10.3390/app9020218
Wang, J., Lan, Y. B., Zhang, H. H., Zhang, Y. L., Wen, S., Yao, W. X., & Deng, J. (2018). Drift and deposition of pesticide applied by UAV on pineapple plants under different meteorological conditions. Int. J. Agric. Biol. Eng., 11(6), 5-12. doi:https://doi.org/10.25165/j.ijabe.20181106.4038
Wang, X. N., He, X. K., Song, J. L., Wang, Z. C., Wang, C. L., Wang, S. L., . . . Meng, Y. (2018). Drift potential of UAV with adjuvants in aerial applications. Int. J. Agric. Biol. Eng., 11(5), 54-58. doi:https://doi.org/10.25165/j.ijabe.20181105.3185
Wolf, T. M., & Caldwell, B. C. (2001). Development of a Canadian spray drift model for the determination of buffer zone distances. Expert Committee on Weeds-Comité d’Experts en Malherbologie.
Xiao, J., Chen, L., Pan, F., Deng, Y., Ding, C., Liao, M., . . . Cao, H. (2020). Application method affects pesticide efficiency and effectiveness in wheat fields. Pest Manag. Sci., 76(4), 1256-1264. doi:https://doi.org/10.1002/ps.5635
Xue, X. Y., Tu, K., Qin, W. C., Lan, Y. B., & Zhang, H. H. (2014). Drift and deposition of ultra-low altitude and low volume application in paddy field. Int. J. Agric. Biol. Eng., 7(4), 23-28. doi:https://doi.org/10.3965/j.ijabe.20140704.003
Yan, X., Zhou, Y., Liu, X., Yang, D., & Yuan, H. (2021). Minimizing occupational exposure to pesticide and increasing control efficacy of pests by unmanned aerial vehicle application on cowpea. Appl. Sci., 11(20), 9579. doi:https://doi.org/10.3390/app11209579
Zhang, H., Qi, L., Wu, Y., Musiu, E. M., Cheng, Z., & Wang, P. (2020). Numerical simulation of airflow field from a six-rotor plant protection drone using lattice Boltzmann method. Biosyst. Eng., 197, 336 – 351. doi:https://doi.org/10.1016/j.biosystemseng.2020.07.018