Comparison of the physical spray efficacy between unmanned helicopter and motorized knapsack sprayer in Thai paddy field

The physical spray efficacy was compared between unmanned helicopter and motorized knapsack sprayer for the application of pesticide in rice paddy field in Thailand. The unmanned helicopter treatment resulted in a significantly higher number of droplet density than that of the motorized knapsack sprayer treatment. In addition, droplet deposition did not differ between the two spraying techniques on the top and panicle positions. However, droplet deposition on the bottom position from the unmanned helicopter was higher than that from the motorized knapsack sprayer. The unmanned helicopter effectively reduced droplet losses to the ground compared with that of the motorized knapsack sprayer. With respect to droplet drift from sprayed area, the distance achieved by the unmanned helicopter was 3 m greater than that by the motorized knapsack sprayer. These results suggest that the efficacy of the unmanned helicopter was similar or greater than that of the motorized knapsack sprayer. Overall, we found that unmanned aerial vehicles improved performance and enabled rational pesticide application rice production in Thailand


Introduction
Rice is an important crop that plays a major role in the economy of Thailand (Office of Agricultural Economics, 2018).However, approximately 50 different species of pests have been reported to damage rice crops in Thailand (Rice Department, 2018).Typically, insecticides to control these pests are applied at a high spray volume using a spray lance fitted onto a motorized hydraulic knapsack sprayer by farmers in Thailand (Pojananuwong et al., 1999(Pojananuwong et al., & 2001)).The efficacy of this control measure depends on the Original Article operator's skill and efforts.In addition, studies have reported that farmers' spraying techniques may pose a safety issue regardless of personal protective equipment and appropriate spray application techniques (Hughes et al., 2008;Nuyttens et al., 2009).The number of reported cases of pesticide poisoning among farmers in Thailand has considerably increased from 2010 to 2017 (Ministry of Public Health, 2018).Moreover, the estimates of operator exposure have been reported for several classes of pesticides and from diverse types of application equipment (Hughes et al., 2008;Nuyttens et al., 2009).These figures reveal the necessity of improved site-specific occupational hygiene.Importantly, the non-uniformity of dermal exposure from different kinds of spray application techniques and spray volumes has been reported (Thongsakul et al., 1999;Machera et al., 2002).All the aforementioned studies have raised concerns about the need to improve spray application techniques.At present, the safety of operators, price of insecticides, lack of laborers, and increasing cost of operating in paddy fields have to be improved on spray application techniques to achieve better results (Wechakit et al., 2009).In recent years, unmanned aerial vehicles (UAVs) has become an interesting alternative application method in several countries in Asia (Xue et al., 2016).At present, UAV's are effective sprayers and play an important role in controlling all sorts of pests, especially in rice (Xue et al., 2008 andQin et al., 2018).In Thailand, some agricultural chemical operators, farmers, and spray service providers are using UAVs in paddy fields.Most UAVs, however, are imported but they were never tested for efficacy.Hence, basic information on good UAVs application practice has to be explored.Therefore, in this study, the number of droplet density on target areas, droplet deposition on rice canopy, droplet losses to the ground, and droplet drift from the sprayed area resulting from unmanned helicopter in paddy field were compared with those from motorized knapsack sprayer to optimize UAVs application techniques in Thailand.

Spray application techniques
Four application treatments were designed for testing (Table 1).The first two applications utilized the FAZER helicopter (Yamaha Corporation, Hamamatsu, Japan) with two fan-type nozzles (XR 110025; Spraying Systems Co., Ltd., Glendale Heights, IL, USA) at a rate of 8 or 16 L/ha, designated as unmanned helicopter 1 (UH1) and UH2, respectively.The other two applications utilized a motorized knapsack sprayer (Maruyama model MS 073D) with a tank capacity of 25 L (Maruyama Co., Ltd., Tokyo, Japan) and an attached spray lance that operated at a rate of 250 or 375 L/ha, designated as motorized knapsack 1 (MK1) and MK2, respectively.deposition on rice canopy, droplet losses to the ground, and droplet drift from sprayed area under working conditions.In this experiment, the results were expected to vary both across the swath width, spray direction and the meteorological conditions.Thus, sampling positions in this experiment were selected to account for this factor as much as possible.To avoid cross-contamination between plots, the sampling site was located 20 m from the edge of each treatment plot (Xinyu et al., 2014 andWang et al., 2017).

Environment monitoring
The ambient temperature and relative humidity at 1, 2, and 3.0 m above the target areas were recorded using the Extech 45160 Humidity, Temperature, and Airflow Meter and the Extech 42270 data logger (Extech Instruments, Waltham, MA, USA).

Sampling methods
Fig. 1 shows how the sampling collectors and deposition, and losses to the ground were arranged in the sprayed area.The sample collectors for measuring the number of droplet density on target areas consisted of a 25 mm × 75 mm of water-sensitive paper (WSP).It was used to evaluate the number of droplet density.The WSPs were adjusted to assist in positioning the papers at a height equivalent to the top and bottom positions of the rice canopy at the tillering and heading stages and fixed vertically on rice panicle at the flowering stage (upwind and downwind positions).The interval of sampling points was 2 m in the rice canopy according to Fig. 1.After spraying, naturally dried WSPs were placed in plastic bags with labels indicating the treatment of spraying.The bags were then tightly sealed and kept in a UV-proof container, to protect the dye and taken to the lab; the number of droplet density on target areas was computed using the image processing software Deposit Scan.Droplet deposition, droplet losses to the ground and droplet drift were determined using the colorimetric method with a tracer dye.Tartrazine, at a concentration of 3%, was chosen for the present experiment, because of its safety, high accuracy and sensitivity with a significant reduce operation time and cost.In addition, it easily can be extracted from Petri dishes and plant cellulose and stable under various light and temperature conditions (Wicke et al., 1999 andPergher, 2001).To determine droplet deposition on rice canopy, the sampling collector consisted of a filter paper (ϕ = 70 mm).Before spraying, a stapler was used to attach the filter papers to rice leaves and panicles in the rice canopy.The filter papers were placed as shown in Figure 1.To determine spray losses to the ground, Petri dishes attached to a metal spike were placed 0.05 m above ground level.The interval of sampling points was 2 m in the sprayed area as shown in Fig. 1.Droplet drift from sprayed area was evaluated by setting up sampling lines downwind where Petri dishes of 90 mm diameter were attached to a metal spike were arranged 1-m apart at the top of the rice canopy up to 20 m apart from the sprayed area (Fig. 2).The collectors were taken from the field 30 min after finishing the application and washed using distilled water.The washing solution for each position was stored for further analysis.In the laboratory, the samples were analyzed using a colorimetric method to determine tartrazine content in the washing solutions.For making the tracer dye spray more accuracy, after UH took off and was hovering 20 m from the sprayed area and stopped spraying 10 m away.The flight height (above the crop surface) was 3 m.

Detection of tracer dye
In all experiments, the samples were analyzed by colorimetry (Jenway model 6051; Spectronic Camspec Ltd., Leeds, UK) to measure the specific absorption of tartrazine at 470 nm.After spraying, the samples were removed carefully with a pair of forceps and each sample was placed in a zip-lock bag and labeled.The samples were spiked in the laboratory with a standard calibration solution.In spite of the high stability of tartrazine, the samples were stored and transported under cool and dark conditions.The tracer was rinsed off the patches using 10 mL of demineralized water in disposable Petri dishes.For analysis of sample, a colorimeter with a micro-plastic cuvette of capacity 3 mL was used.The readings are expressed in µg/L.The amount of tracer contained in the sample was Asian J Agric & Biol.2021(4).calculated for each measurement target.An allowance was made for various correction factors, such as dilution, measurement range and the volume of the absorbent liquid.The results of droplet deposition are expressed in µg/unit area.

Statistical analysis
Before the analysis of significant difference, the number of droplet density on target areas, total droplet deposition, and losses to the ground, were transformed by using log (x + 1).Significant differences were determined using an analysis of variance (ANOVA) and Tukey's test at a significance level of 95% with SPSS v. 22.0 (SPSS, Inc., IBM, Chicago, IL, USA).

Environment during field performance
The average wind speed, ambient temperature and relative humidity of the experiment are shown in Table 2.

Number of droplet density on target areas, the total droplet deposition and losses to the ground
The number of droplet density on target areas from both unmanned helicopter treatments was 46.8-92.6 droplets/cm 2 , higher than that of the motorized knapsack sprayer treatments (29.8-78.6 droplets/cm 2 ) (Table 3).The spraying technique had a major effect on the number of droplet density.Previous studies have reported that effective treatments for insecticide or preemergence herbicide applications to control insect pests and weeds require at least 20-30 droplets/cm 2 , for contact post-emergence herbicide applications require at least 30-40 droplets/cm 2 , and for fungicide applications require at least 50-70 droplets/cm 2 , respectively (Matthews et al., 2014).In addition, Ebert et al. (1999) reported that the number of spray deposits only needs to reach a certain threshold to achieve sufficient control efficacy.Based on these reports, the number of deposits from both unmanned helicopter treatment was sufficient for all types of pesticide.The total droplet deposition with the unmanned helicopters on the top position and panicle position was not significantly different from that of the motorized knapsack sprayers.Although UH2, MK1, and MK2 delivered a higher spray volume than that of UH1, the volume of dye deposition was not significantly different from that of UH1, which operated at the lowest spray volume of 8 L/ha.On the contrary, the total deposition of the unmanned helicopters on the bottom position was significantly different from that of the motorized knapsack sprayers (Table 4).This result suggested that the air stream from downward direction below the rotors may increase droplet penetration when the unmanned helicopter is operated for very low volume spraying.In addition, the air stream generated by the rotor wings has perturbed the rice plants, as the droplets often penetrated into the canopy and reached to the target areas.Although the motorized knapsack sprayers delivered 90% more spray volume than the unmanned helicopters, they did not produce significantly higher densities and deposits than that of the unmanned helicopters, which operated at lower spray volumes of 8 and 16 L/ha.Values followed by the same letter in the column do not differ statistically (p < 0.05; Tukey's Test).UH: unmanned helicopter, MK: motorized knapsack.
In addition, most of the smaller droplets are free in the air, and more easily reach the lower canopy of the crop.Therefore, smaller droplets may have better penetration than larger droplets for the ground spraying equipment.This result agrees with Qin et al. (2018) and Wang et al. (2019) who found that the impact of UAV spraying on the distribution and deposition of droplets on canopy of wheat was quite significant especially during the late growing stages when the leaf area index was the largest, the droplets coverage by routine ground sprayer was distinctly lower than that realized by UAV.The unmanned helicopter techniques showed relatively lower dye deposition and, correspondingly, the two motorized knapsack techniques MK1 and MK2 resulted in higher losses to the ground by more than 35%-39%, 32%-41%, and 25%-44% on rice at the tillering, heading, and flowering stages, respectively (Table 3).
As reported in other studies, it is likely that high-and medium-volume sprayings lead to increased run-off (Sánchez-Hermosilla et al., 2012;Rincón et al., 2017 andWang et al., 2019).Spraying with coarse droplet resulted in high losses to the ground inside the field when compared with that of spraying with fine droplet (Derksen et al., 2007).Our results showed a similar trend to that of the previous studies.

Droplet drift deposition from sprayed area
Throughout the experiment, the height of the unmanned helicopters was 3 m above the top of the crops, wind speed was less than 0.3 m/s, average ambient temperature was 24 °C, and average relative humidity was 82%.Table 5 shows that the droplet drift was within 8 m from the sprayed area.On the Petri dishes that were placed 9-20 m from the sprayed areas, the droplet drift was negligible.We considered that because of the suitable weather conditions and the vertical spray direction provided by the unmanned helicopter, droplet drift from sprayed area was reduced.This, coupled with the low wind speed of less than 0.3 m/s, mild ambient temperature of 24 ºC, and relative humidity of 82%-all of which are favorable weather conditions-resulted in droplets spreading not beyond 8 m outside the targeted area; the distance achieved by the unmanned helicopters was only 3 m greater than that of the motorized knapsack sprayers.The similar result was described by Kamthonsiriwimol et al. (2020) that spraying drift by using UAV in the paddy field was greater than knapsack sprayer.According to the UAV spraying, the droplet size was smaller than the one of knapsack sprayer.Therefore, the drift ability of UAV spraying was higher than conventional spraying by using knapsack sprayer.These findings correspond to those of Xinyu et al. (0214) and Wang et al. (2017), who found that when spraying with a UAV at a height of 3 m above the plant with a wind flow speed of less than 1 m/s, the spread of droplets was only found within 10 m of the last row where fungicides were sprayed.

Conclusion
The number of droplet density on target areas and droplet deposition on the bottom position from the unmanned helicopters were higher than those from the motorized knapsack sprayers, which is currently the conventional spraying method in Thailand.Furthermore, the unmanned helicopters effectively reduced droplet losses to the ground with those of the conventional method.The efficacy of the unmanned helicopter was similar or greater than that of the motorized knapsack sprayer.Overall, UAVs displayed improved performance, and their use is a rational strategy for pesticide application in Thailand.
Experiment site data collection, processing, and analysis Field experiments were conducted in Suphanburi, Thailand.The experiments were performed using a randomized complete block design, with four spray application techniques and five replicates per technique.Each plot size had an area of 50 m × 24 m.Field studies were conducted on rice at 45, 70, and 90 days after sowing (represented the most frequency to spray insecticide and fungicide in the field) to compare the number of droplet density on target areas, droplet Pruetthichat Punyawattoe et al 3/8 Asian J Agric & Biol.2021(4).