https://orcid.org/0000-0002-0518-7631
Abstract
Cranberry fruitworm (Acrobasis vaccinii Riley (Lepidoptera: Pyralidae)) and blackheaded fireworm (Rhopobota naevana Hubner (Lepidoptera: Tortricidae)) threaten cranberry production annually by causing significant fruit damage. Up to four pesticide applications are made each year to control these insects, which are costly to producers and elevate pesticide residues in fruit. Pheromone-based mating disruption technology can provide control of these pests in cranberry production, with the potential to minimize, or eliminate, pesticide applications. In 2016, an uncrewed aerial vehicle (UAV) was investigated to apply a thick paraffin emulsion containing insect sex pheromones. Traditional agricultural equipment is not capable of applying the paraffin emulsion to cranberry beds due to the product’s viscous, paste-like consistency. The first objective of this study was to retrofit an UAV (octocopter) with a novel extrusion device that had been engineered to deliver the pheromone-loaded paraffin at regular intervals during flight. The second objective was to confirm adequate distribution of the pheromones by measuring the mating disruption efficacy by monitoring male moth trap catches. The UAV was able to fly autonomously along a prescribed itinerary, deploying the paraffin product uniformly; however, the increased mass of the retrofitted UAV limited flight times to ~12 min. The number of male cranberry fruitworm and blackheaded fireworm moths caught in lure-baited traps were reduced in the paraffin-treated beds compared with untreated beds, indicating adequate distribution of the pheromones. The UAV-applied pheromones concept could be developed into a production scale application method in the future, although issues of battery life and lifting capacity will need to be resolved.
Uncrewed Aerial Vehicles (UAV; drones) are increasingly generating substantial interest within the agricultural community (Freeman and Freeland 2015). Advancements in electronic control technologies, increased battery life, and Global Navigation Satellite Systems (GNSS) have allowed UAVs to enter a paradigm of mainstream use. Incorporation of UAVs into agricultural production systems has been centered around high-resolution remote sensing data collection with relatively small UAVs (Freeman and Freeland 2015, Pajares 2015, Gabriel et al. 2017, Aasen et al. 2018, Feng et al. 2020). However, using a slightly larger UAV platform provides the potential for application of products to crops during the growing season. For example, Giles and Billing (2015) utilized a petrol-powered single rotor helicopter aircraft to apply sprayable pesticides to a vineyard. Results showed that the UAV platform tested could achieve working rates of 2–5 ha/h. Martinez-Guanter et al. (2020) conducted a similar study with a small battery-powered UAV applying liquid agrochemicals to olive and citrus orchards. Working rates for this UAV platform were found to be 5.4 ha/h; however, the authors speculated that working rates could be doubled by increasing the flight speed of the UAV platform in suitable weather conditions. UAVs often demonstrate the ability to apply low-volume inputs to crops efficiently. They also provide the added advantage of applying these products easily where terrain is challenging for people and/or machines to reach.
North American cranberries (Vaccinium macrocarpon) are a high-revenue, input-intensive crop with very specialized machinery needed to manage the production system (Eck 1990). In Wisconsin, cranberries are generally grown in beds engineered for water management, specifically the use of contoured sand and raised dikes around the perimeter to facilitate efficient water movement as well as equipment access. Fertilizer and pesticide applications are completed using a cantilevered boom sprayer/spreader combination in order to apply these products to the beds without driving over the crop (Fig. 1). These machines are generally customized to the bed width of the specific marsh on which they are used. Maneuvering these pieces of equipment around the marsh for various applications of crop inputs can be cumbersome and time consuming due to the cantilevered design. Having the ability to provide timely application of low-volume crop inputs without using the boom applicators would give cranberry growers an additional tool for managing their crop. In eastern cranberry bogs, there is further interest in UAVs, given the greater variability in bed shape, hydrology of the bog, and spatial distribution of the beds.
Cranberry marsh combination sprayer/spreader application equipment. Cantilevered design allows farmers to apply fertilizers and pesticides across the entire bed without driving on the plants. Dikes are constructed on the perimeter of the beds to carry equipment and provide water control structures.
Two prominent cranberry pests, cranberry fruitworm, Acrobasis vaccinia Riley (Lepidoptera: Pyralidae) and blackheaded fireworm, Rhopobota naevana Hübner (Lepidoptera: Tortricidae) invade the beds annually. These herbivores damage the leaves and fruit of the plants and are considered some of the most concerning pests for cranberry growers in Wisconsin (Chasen and Steffan 2016). A mating disruption system has been tested as a means to control these pests. This system used a flowable paraffin emulsion as the carrier for the insect pheromones and was shown to control both cranberry fruitworm and blackheaded fireworm (Steffan et al. 2017). By preempting mate-finding in these moths, it was able to reduce caterpillar densities, thereby minimize pesticide applications and providing evidence that mating disruption systems could become a viable tool for U.S. cranberries. However, the early application methods involved manual applications of the paraffin, and this was prohibitively labor-intensive to adequately distribute over the cranberry beds in a timely fashion. This is due to the viscosity of the product which prohibited it from being applied as a liquid product through typical agricultural sprayer, and yet not solid enough to be distributed by a typical agricultural spreader. Additionally, a delivery system was required to eliminate the need to enter the beds to apply the product and have the capability of extruding the highly viscous matrix at regular intervals to provide adequate coverage (Steffan et al. 2017). Thus, the objectives for this research project were to build upon the 2017 work as follows: 1) develop an UAV system capable of efficiently delivering a wax-based mating disruption product to cranberry beds, and 2) verify that the delivered product is in sufficient quantities to cause mating disruption during moth flight.
Materials and Methods
Extrusion System and UAV Autonomy
We set out to design an extrusion system that dropped dollops of the mating disruption product at regular intervals to ensure uniformity of coverage of the pheromones within the bed. Based on previous mating disruption work in the cranberry system (Steffan et al. 2017), design requirements for this device were as follows:
Have the capability of extruding the paraffin emulsion loaded with pheromones and produce 1-g dollops.
Drop the dollops 3.04 ± 1.5-m intervals within the bed.
Provide enough product capacity to treat 1/3 of a 0.37-ha cranberry bed (approximately 1,230 m2).
Total weight of the mechanism and mating disruption product should not exceed 6 kg.
The UAV chosen for this project was the Spreading Wings S1000 (DJI, Shenzhen, China). This platform provided sufficient lifting capacity, advertised as 6–13 kg, for the addition of the application mechanism. Power was supplied to the UAV via two 16,000 mAh lithium ion batteries (PLU15-160006) connected in parallel. These batteries advertised sufficient power to achieve ~15 min of flight time when considering the additional payload. Control of the UAV was achieved using a hand-held controller (Futaba T14SG) with a line-of-sight communication range of ~1,500 m. This manual control was coupled with DJI Ground Station software (DJI, Shenzhen, China), which allowed for automated flight planning and parallel pass flight patterns when connected to the flight controller. A BlueTooth radio (DJI LK24-BT, DJI, Shenzhen, China; 2.4 MHz) was used to connect the flight control software to the UAV during flight. The A2 Flight Controller (DJI, Shenzhen, China) was attached to the UAV providing control of the aircraft and guidance via Wide Area Augmentation System (WAAS) corrected GNSS. Flights were conducted at ~4 m above the cranberry bed to limit wind influence on the dollops as they were applied.
Several design variations for the mating disruption application system were considered. Aside from the design specifications listed above, this system presented an interesting problem in that the high viscosity of the wax emulsion eliminated the possibility of using typical pumps to move the product and extrude 1-g dollops. All commercial pumps available would far exceed the weight restriction required to mount the system to the UAV. The design team considered alternative mechanisms to achieve material movement and consistent dollop extrusion. One design considered was an assisted gravity fed system implementing two Archimedes’ screws (Fig. 2). Construction of this mechanism was completed using off-the-shelf components including a funnel and polyvinyl chloride (PVC) fittings. The PVC fittings and funnel were machined to achieve a sliding fit with the screw element. Hay anchors (Item # 227145, Gempler’s, Janesville, WI) were used as the screw elements and modified to attach to electric motors to provide the rotation.
Initial mating disruption product application system design generating material movement with two Archimedes’ screws. (A) Bulk product container with vertical Archimedes’ screw. (B) horizontal screw for wax emulsion extrusion and application. Dollop production was achieved with this system; however, limited material capacity and lack of dollop drop rate consistency removed this approach as a viable solution.
The final mating disruption product application system design was based on a piston and cylinder configuration. A 38.1-cm length of clear 10.2-cm internal diameter PVC tube was used as the main body of the application system (Fig. 3A). This tube housed a piston made from Delrin plastic (Fig. 3B) machined to a sliding fit within the tube. Two o-rings were installed on the flat section of the piston to prevent the paraffin emulsion ingress to the rear of the tube and to wipe the emulsion from the sides of the tube. The front of the piston was formed to closely mate with the corresponding nozzle, which had a diameter of 15 mm (Fig. 3C) in order to fully empty the tube prior to refilling. The plunger was advanced using a ball-screw linear actuator (17RNA150-4-2516, Thomson Linear). This actuator was capable of producing 23 kg of force when powered at 12 VDC.
Laboratory testing of the mating disruption product application system. The PVC tube (a), piston (b), and nozzle (c) provided the capability to extrude dollops of the wax-based/pheromone emulsion at regular intervals during flight.
Control of the extrusion system was achieved using a microcontroller (Uno, Arduino) coupled with a motor controller breakout board (L6470 Auto Driver, Sparkfun Electronics) (Fig. 4B). A digital input/output pin on the microcontroller was connected to the receiver of the hand-held remote control, which allowed for actuation of the system via a switch on the remote control. Once actuated, the microcontroller would move the ball-screw linear actuator forward approximately 13 mm at step intervals of 1 s. The control and distribution systems were powered using a 14.8 VDC 1300 mAh lithium ion battery pack (XT60 Lumenier) separate from the batteries powering the UAV.
Spreading Wings S1000 UAV retrofitted for mating disruption product deployment. Standard industry flight controller, batteries, and propellers were used for this proof-of-concept testing. Camera gimbal mounting locations were adapted to carry the electronic control system as well as the mating disruption product distribution system.
Loading the mating disruption product into the application system consisted of manually filling the tube with a spatula after removing the motor and piston assembly. Tape was placed over the nozzle so that the product did not leak out during loading. Once the tube was filled the product was agitated with an offset paddle paint stirring device (447, Warner Mfg. Co.) connected to a battery powered drill to remove any air pockets created during filling. Once agitated, the remaining space in the tube was filled with the product and the piston was re-installed.
A plastic adapter plate (Fig. 4A) was fabricated to connect to the standard gimbal mounting location on the UAV. The control system and auxiliary battery were mounted to the top of the adapter plate and the distribution system was connected via two hook and loop straps (HF-25R5/BK, Envisioned Products, Inc) to make removal and re-loading less time consuming.
Laboratory tests were conducted to tune the frequency and duty cycle inputs to the application system to produce dollops of sufficient size and at the proper drop frequency. Once the tuning of the system was complete, dollops were caught on 51 × 51-mm paper cards and weighed on a laboratory scale (Scout Pro SP601, Ohaus Corporation) to ensure the dollop size was within acceptable limits.
UAV Deployment of a Mating Disruption Product
Previous work had indicated that a flowable paraffin emulsion (a proprietary blend of paraffin, solvents, and pheromones) represented a proven means of dispersing, protecting, and slowly volatilizing the sex pheromones of moth species, for the purpose of disrupting moth mating in U.S. cranberries (Steffan et al. 2017). Given the success of past proof-of-concept studies, we designed the current study using just two commercial cranberry marshes, referred to as Marsh A and B, located in Northern Wisconsin. Testing at only two sites allowed us to focus resources on developing and deploying the drone application, rather than re-testing the efficacy of the material itself. Each marsh was managed under standard grower practice, and the mating disruption treatment was applied as an addition to such practice at each marsh; thus, each treated block could be contrasted with the corresponding standard practice. Based on past work, ~300 ‘dollops’ per acre represented an adequate point-source density (Steffan et al. 2017), so the UAV system was programmed to uniformly deliver this density of point-sources.
Marshes were selected based on grower interest in the project. Beds within the marshes were chosen based on pest history and bed location within the field. Specifically, mating disruption treated beds were chosen to be at the marsh edge and to the north and/or east so that predominant winds (out of the southwest, typically) do not cross-contaminate untreated beds. By minimizing the amount of pheromone that would be distributed by wind throughout the marsh, we were better able to compare the treatment effect to the grower ‘control’ (i.e., standard grower practice). Control beds were chosen to have similar pest history to treatment beds while being at least 200 m from the treatment bed. One bed at each site was chosen for pheromone treatment. The total treated acreage between the two sites was ~2.4 ha.
To monitor the effectiveness of the mating disruption product, two sets of Pherocon II traps (Great Lakes IPM, Vestaburg, MI) containing lures for either CFW or BHFW were placed in each treatment and control bed, according to standard monitoring protocols in cranberries. Two traps each for CFW and BHFW were placed during deployment of the mating disruption product. Traps were checked weekly for 6-wk postapplication to count the number of male moths caught. Under successful mating disruption, we expect to find fewer moths in traps placed in pheromone-treated beds, given that the pheromone plumes from the treatment should mask the plume from the lure.
Results and Discussion
An aerial application system was designed by retrofitting an UAV with an extrusion device. This device was mounted on the UAV and dispensed a pheromone-loaded paraffin emulsion to commercial cranberries. A preprogramed flight path allowed the UAV to fly autonomously and extrude the paraffin emulsion onto the cranberry canopy. The aerial application system worked well in distributing the pheromones uniformly, providing a means to deploy a mating disruption system for CFW and BHFW.
Initial design iterations of the mating disruption application systems centered around the Archimedes’ screw concept, which showed promise as a workable solution; however, this had characteristics not ideal for this system. The gravity assisted nature of the funnel with a vertical screw at the center was not able to feed the paraffin emulsion consistently to the horizontal extruding screw. Air pockets and bridging of the wax emulsion would occur at random intervals during operation. This irregular delivery of the material led to inconsistent dollop size and drop interval. The capacity of the funnel was also identified to be insufficient to carry enough wax emulsion to accommodate the intended flight times and application rate.
The final piston/cylinder application system construction and implementation was successful. The ball-screw actuator provided sufficient force to move the piston within the tube to extrude dollops of the mating disruption product. The Arduino-based control system also succeeded in controlling the application system. Laboratory tuning of the control system resulted in an input frequency to the actuator of 0.33 Hz and a duty cycle of 33% to achieve the required dollop drop rate of 3.04 ± 1.5 m. Connection to the hand-held radio provided the ability to engage and disengage the system remotely. The timer-based control algorithm allowed the system on/off frequency and duty cycle to be adjusted so that dollops were dropped at an acceptable size and rate. Benchtop testing of the application system with the wax emulsion showed the performance was sufficient (n = 7, M = 1.64 g, SD = 0.44 g) to implement the system in field testing. The wax emulsion used for testing was not loaded with pheromones which we assumed would make the product slightly more viscous than the wax emulsion with the pheromones included. Thus, a mean dollop of 1.64 g dropping with only gravitational forces acting on it was considered acceptable.
Prior to taking the UAV-based application system to field sites several test flights were conducted. These tests were completed with and without the mating disruption product payload. The UAV platform used in this study was not specifically designed with product application in mind. The flight parameters of this project required the UAV to fly low to the ground, parallel passes, and land in a small area ~3.7 m wide. Testing of the flight controller and Ground Control software showed the UAV could maintain an altitude of 4 m above ground level during flight. One point of concern was as the UAV transitioned out of a hover and accelerated forward it lost altitude briefly. Recovery from this loss of altitude occurred after ~22.9 m. Pass-to-pass distance during flight was sufficiently equal to move forward with the project. The flight control system installed on the UAV was equipped with automatic take-off and landing capability. However, this system was not precise enough to ensure proper starting altitude nor ensure the UAV would land on the 3.7-m wide dike upon returning to its starting position. Thus, our process for conducting an application flight consisted of take-off and landing conducted by a certified remote pilot prior to engaging the application system and automatic guidance of the UAV (Fig. 5).
(A) Remote pilot operated take-off for applying the mating disruption product via UAV, and (B) automated guidance of the UAV once in the air and at the proper altitude. Note the extrusion device mounted on the underside. Landing procedure was conducted by the remote pilot as well to ensure a safe landing on the dike surrounding the cranberry bed.
Field testing of the UAV-based mating disruption application system was completed at two locations in northern Wisconsin. The system could fly with the additional 6 kg of payload using manual control by the remote pilot for take-off and landing. As stated above, maintaining an altitude of 4 m above ground level proved to be a concern. At this low altitude, the time for the remote pilot to take control of the aircraft and correct altitude or flight path errors before an unplanned landing occurred was minimal. On two occasions during the field testing the aircraft flight controller lost communication with the Ground Station software. This issue caused a hard landing in the bed on one occasion and an inverted crash landing in the bed on another. Neither of these mishaps damaged the aircraft sufficiently to stop testing. Any failures in landing gear, frame, or propeller components were noted and these parts were replaced before testing resumed.
A notable shortcoming of the UAV used in this testing was battery life. The nature of this application system required slow flight speed so that the dollops could drop straight down and remain intact as much as possible. The forward flight speed was ~4.8 km/h. Due to the slow speed and the additional weight of the mating disruption application system, the battery only had power for two parallel passes over the 122-m long beds. This corresponded closely with the amount of the mating disruption product in one load. Thus, after every two passes, the aircraft would land, the batteries would be changed, and the mating disruption application system would be re-filled. The beds tested required 8–10 total passes of the system to receive adequate coverage. Recharge time for each pair of batteries was ~1 h. Total application time at each location was 6–8 h including battery recharge time yielding 0.05–0.04 ha/h working rates. Other studies applying liquid fertilizers and pesticides have reported working rates ranging from 2.0 to 5.4 ha/h (Giles and Billing 2015, Martinez-Guanter et al. 2020). Application speed when applying liquid products is generally higher than 4.8 km/h and Giles and Billing (2015) used a petrol-powered aircraft instead of battery power. Having longer flight times, reducing the frequency of battery changes, and heavier payload capacity would make this UAV application system a more viable solution in the future.
Figure 6 shows two examples of mating disruption product dollops in the cranberry beds. Dollop size during field testing was similar to that in laboratory testing based on visual observations. The pheromone loaded wax emulsion used during field testing seemed to have a higher viscosity than the product used in laboratory testing as the dollops did not drop as quickly. Distances between dollops ranged from 3.7 to 6.1 m. This resulted in a lower dollop density (per unit area) within the bed. Having a consistent mating disruption product that has similar properties between batches would greatly aide in achieving proper application with the UAV-based application system.
As applied mating disruption wax emulsion product to production cranberry beds via UAV. The dollops of mating disruption product were slightly larger than the 1 g goal. Dropping this product from a height of 4 m above the crop canopy allowed some dollops to fracture (A) rather than remain whole (B) which will shorten the pheromone emitting time (SAS, unpublished data).
Dollop shape was another factor qualitatively assessed within this study. Previous research conducted by Steffan et al. (2017) showed the efficacy of the pheromone mating disruption wax emulsion. In this work, the product was applied by hand and provided a very consistent dollop size and shape. However, dropping the dollops from an altitude of ~4 m with the UAV tended to fracture the dollop into smaller pieces (Fig. 6A). Pheromone discharge rate from these dollops is affected by several factors including weather conditions and surface area-to-volume ratios of the dollops (S. A. Steffan, unpublished data). A more regularly shaped dollop, as in Fig. 6B, will emit the pheromone more uniformly and for a longer time due to its lower surface area-to-volume ratio. We speculate that the addition of a mechanism to cut the dollops off the nozzle and possibly adding shielding to interrupt wind and propeller wash would allow the dollops to fall in a more regular shape. Allowing gravitational forces to determine when the dollop dropped from the nozzle allowed for stretching and fracturing of the dollops.
Figure 7 shows the number of male moths caught in the Pherocon II traps in both the pheromone-treated beds and the control (grower standard practice) beds at the two marshes where this study was conducted. Cranberry fruitworm trap-capture numbers (Fig. 7A) were greatly reduced in the pheromone treated bed at Marsh A while there was marginal reduction at Marsh B. These results specifically show that we were able to effectively use the drone to deliver the product in quantities sufficient to disrupt mating. Blackheaded fireworm trap-capture numbers (Fig. 7B) were very low during this growing season at the test marshes and were not sufficient to draw any meaningful conclusions regarding the pheromone treatment. While being a top pest in cranberry production, blackeaded fireworm populations are notoriously spotty (for example, see (Guédot and Lippert 2019, Guédot et al. 2020).
Pheromone-baited trap catches of adult male moths of the cranberry fruitworm and blackheaded fireworm in the mating disruption-treated (MD) and control cranberry beds. Treated beds showed lower trap catches than the control beds; however, replication was not sufficient for statistical analysis (replication was limited by issues with UAV battery life).
Considering that our previous work (Steffan et al. 2017) had already demonstrated the efficacy of this particular mating disruption product, the current research focused on engineering a new way to mechanize the deployment of the pheromone product. To these ends, we fabricated a novel device that was able to extrude the pheromone-loaded paraffin while attached to an in-flight UAV. When fully loaded, the extruder added significant weight to the UAV. Despite this extra payload, the UAV was able to fly autonomously while self-correcting, based on a preset flight itinerary. Collectively, this engineering success allowed us to assess the viability of applying a paraffin emulsion via a retrofitted UAV. We found that battery life was insufficient, and that the time needed to reload the extrusion cylinder made the whole process prohibitively time-intensive to cover large acreages. If the cylinders could be preloaded with the paraffin carrier, and if enough batteries were on-hand, a farmer could potentially cover significant acreage efficiently. Having a drone that can fly autonomously, ‘knowing’ where it is spatially, and deliver a product to a prescribed mark, should hold significant intrinsic value for a farmer, and future refinements to such systems will strike a balance between accuracy, precision, and efficiency.
Conclusions
Battery-powered UAVs represent a remarkably powerful means of autonomously and remotely delivering pest management technologies, but the trade-off between payload and battery-life needs to be weighed (both figuratively and literally). For mating disruption programs, current battery-powered UAVs do not contain enough power density to accommodate the additional payload and slower air speed necessitated by the paraffin carrier of the pheromones. The weight of the pheromone load, itself, may be minimal, but the carrier of the pheromones presents challenges for the UAV’s payload. Future work may need to focus on lower-density carriers and/or larger, petrol-powered UAVs. As battery technology advances, or larger, petrol-powered UAVs become more cost-efficient to own and operate, the UAV-based mating disruption application system could become a viable program for cranberry production. Indeed, this sort of product-delivery system may become broadly useful for a wide variety of crop protection technologies in virtually any agricultural system. A short video of the system described in this manuscript is included as Supp material (online only).
Supplementary Data
Supplementary data are available at Journal of Economic Entomology online.
Acknowledgments
We thank the Wisconsin Cranberry Growers Association (grant 16-03 awarded to Luck and Steffan) for providing funding support for this research project. We would also like to acknowledge the efforts of David Geisler of Rapid Imaging Solutions, who served as the licensed remote pilot and Federal Aviation Administration Section 333 exemption holder for this project.
