SPRAYING QUALITY USING UNMANNED AERIAL VEHICLE IN CITRUS

This study aimed to evaluate the effect of the operating flight height of an unmanned aerial vehicle with different spray tip rotations on the parameters of droplet coverage, density and volumetric median diameter of the sprayed droplets, in acid lime culture ´Tahiti’. The experiment was performed in randomized blocks design, with repetition within the block, in a factorial system (3x2), with four repetitions. Three planting lines were selected for each treatment. The two outer lines were used as borders and the central one for evaluation. In each treatment, one plant was selected from the planting line and four water-sensitive papers were fixed in the middle third of the outermost portion of the canopy.A solution containing water and a drift reducing adjuvant was prepared to analyze the quality of the spray. A significant difference was found for the variables of spray tip rotation and flight height with no interaction between them. This result indicates that these variables act independently on the volumetric median diameter values. Furthermore, the spray heights of 3.0 m and 4.0 m have a notable effect on the amount of deposition. However, their influence on droplet size is negligible.


INTRODUCTION
The genus Citrus has several economically important fruits that are cultivated worldwide for their high nutritional value (HAO, 2019). In Brazil, Citrus are amongst the main cultivated fruit trees (KIST et al., 2018). They present great value for the country's economy, promoting socioeconomic growth and generating numerous direct and indirect jobs (NEVES & TROMBIN, 2017).
The main types of citrus grown in Brazil are oranges, tangerines, acid limes, and lemons. Within the genus, the acid lime (Citrus latifolia Tanaka), popularly known as Tahiti lime, has been arousing interest for the expansion of commercial plantations. Brazil has an outstanding production of the fruit, with 92% of its production destined for the domestic market, as fresh fruit, mainly due to its availability throughout the year and good appearance (KIST et al., 2017).
Despite the high production of the 'Tahiti' acid lime, there are many challenges in Brazilian citriculture. The most prominent challenge is the phytosanitary problem, causing the reduction of productivity, decrease of citrus fruits quality, and increase in the cost of production (ALMEIDA et al., 2018). Currently, chemical spraying is considered the most effective method for prevention and control of pests and diseases (LAN & CHEN, 2018) and can be performed manually or mechanically, through ground or aerial spraying (QIN et al., 2016).
According to Tang et al. (2018), the use of unmanned aerial vehicles (UAV) for plant protection has increased in recent years, due to the ease of operation in areas of difficult access. In addition, UAV spraying can reduce input costs by up to 80%, optimizing resources and applying pesticides at the appropriate time and place, providing lower environmental impacts compared to other spraying methods (ANDRADE et al., 2018). However, the spraying performance and operational parameters of unmanned aerial vehicles should be evaluated in order to elucidate the effects on the distribution of sprayed droplets and to achieve the expected efficiency (MENG et al., 2020). Qin et al. (2018) evidenced that spray height influences droplet distribution in wheat. Trials conducted by Lou et al. (2018) reported that uniformity, coverage rate, and droplet deposition varied as a function of flight height in cotton. Tang et al. (2018) evaluating the effects of operating height and citrus canopy shape demonstrated that at 1.2 meters from the citrus canopy there was better spray performance. Chen et al. (2020) demonstrated that for rice crops, the deposition distribution and droplet penetration were influenced by droplet size (CABRAL & VITÓRIA, 2022).
There is little research done on aerial spraying and the ideal operational parameters for citrus trees under Brazilian conditions. Therefore, the objective of this work was to evaluate the effect of the operation height of an unmanned aerial vehicle with different nozzle rotations on the parameters of coverage rate, droplet density and the volumetric median diameter of the sprayed droplets, in the cultivation of the 'Tahiti' acid lime.

MATERIAL AND METHODS
The experiment was conducted in November 2020, at Bello Fruit company farm, Fazenda Rio Preto, located in São Mateus -Espírito Santo (18°49'34.1"S / 39°54'11.2"W), Brazil. According to the Köppen classification, the region's climate is considered hot and humid, type Aw, with a rainy summer season and a dry winter season (ALVARES et al., 2013). The experimental area is a Tahiti lime plantation grafted on the trifoliata 'Flying Dragon' with a spacing of 6.0 m between rows and 2.0 m between plants, and age of 7 years. In the experiment, the planting rows containing trees of 2.5 m in height were selected.
The experiment was performed in randomized blocks design, with repetition within the block, in a factorial system (3x2), with four repetitions. Three flight operational heights in relation to the lime tree canopy (2.0 m; 3.0 m; and 4.0 m) and two spray nozzle rotations (7,500 rpm and 10,500 rpm) were used. The overflight of the UAV was 60.0 m with the direction aligned in the center of the citrus canopy.
The UAV (Joyance®, model JT-5) used in the experiment had a tank capacity of five liters and flight autonomy of 10 to 25 minutes. The drone had six rotors and two centrifugal spray nozzles.

SPRAYING QUALITY USING UNMANNED AERIAL VEHICLE IN CITRUS
The UAV could flight to 12 m s -1 , and its operation speed was from 0 to 8 m.s -1 .
Three planting rows were selected for each treatment in order to minimize experimental errors due to spray drift. The two outer rows were used as borders and the central one was used for evaluation.
In each treatment, one plant was selected from the planting line, and four water-sensitive papers (76×26 mm) were attached to the middle third of the outermost portion of the crown, two in the direction of the row and two between the planting line.
The quality of the spray was analyzed using a solution containing water and drift-reducing adjuvant (Helper Air©) prepared with a spray volume of 10 L.ha -1 . The flight speed was 6 m.s -1 . Nozzle rotation was measured with optical digital tachometer (Akrom Kr98 © , model laser-sighted).
The applications were performed according to the methodology described by ISO 22866 (INTERNATIONAL ORGANIZATION FOR STANDARDIZATION, 2005). Therefore, the temperature should be between 5 °C and 35 °C, the wind speed should have a maximum of 10% of the measurements below 1.0 m.s -1 and the wind directionshould be within a limit of 90° ± 30° in relation to the spraying line (Table 1). Thus, the ideal wind direction for applications should be eastsoutheast (112.5°) and may be between east and southeast (90° and 135°). During the experiment, the environmental conditions were monitored using a portable digital anemometer (Instrutherm, model AD-250 © ) and a digital thermo-hygrometer (Instrutemp, model SH-122 © ).
The DropScope Wireless © , a water sensitive paper reader, was used. It is composed of a software, a web area for viewing and sharing analyses, and a wireless digital microscope with a digital image sensor with over 2500 dpi. After each treatment and label drying, the impacts on each water sensitive paper label were quantified and characterized. The following parameters were evaluated: volumetric median diameter (VMD, µm), droplet density (droplets.cm -2 ) and coverage rate (%).
The Levene and Shapiro-Wilk tests were applied to verify the homogeneity and normality of the residues. Data were transformed when necessary and variance analysis was performed. The Tukey's test was applied to identify the difference between treatments, when appropriate. All tests were performed with the Statistical Analysis System Software (SAS 9.1), considering a 5% significance level.

RESULTS AND DISCUSSION
The effects of operational flight height at two spray nozzles rotation levels on application coverage are shown in Table 2. At 10,500 rpm, no statistical difference in the coverage rate as a function of height was found. This result may have occurred because the higher the rotation, the greater the number of droplet fragmentation and, consequently, the greater their coverage on the  Matthews (1984). In most cases, the spray volume applied by UAV is very low (QIN et al., 2016;WANG et al., 2019), which requires droplet size reduction to optimize crop coverage. However, different droplet sizes can be obtained by changing the rotation of the centrifugal nozzles (QINGQING et al., 2017;RHIND, 2000).
For 7,500 rpm, the flight height of four meters showed greater target coverage. Although lower rotations produce larger droplets and, consequently, less coverage. This treatment shows that the higher height leads to a greater range of spray and, thus, more drops to cover the target due to the downwash effect (ZHANG et al., 2021). Xue et al. (2014) evidenced similar result in their study. Tang et al. (2018) concluded that the spray extension is narrower at low operating heights, thus providing less deposition and coverage of the droplets on the target.
Ideal coverage of droplets in spraying differs according to the type of target to be reached. For insect control, due to greater movement and small number of targets, large drops are recommended (BAESSO et al., 2014). Products that need good coverage, such as protective fungicides, require finer droplets to obtain effective control (CUNHA et al., 2006). According to Contiero et al. (2018), systemic products can be applied in larger droplets, since they are less affected by drift, and ensure target control.
The F test values and the coefficient of variation of the droplet density parameter are presented in Table 2. Most of the variation sources showed no significant differences, except for rotation. This result indicates that the spray nozzles rotation influences the droplet density parameters.
The operating parameters showed no interaction. Therefore, the spray nozzles rotation did not vary with the different flight heights (Table 3).
When analyzing spray nozzles rotation as a function of droplet density, we observed that the rotation of 10,500 rpm provided higher values ( Figure 1A). This result indicates that the higher the rotation, the greater the fragmentation of droplets and the greater the number of droplets per area. The results of droplet density found are consolidated by Matthews et al. (2016), which pointed out a reduction in droplet density in applications with coarser droplets.  The droplet density parameter was not statistically influenced by flight height ( Figure 1B). Nevertheless, trials conducted by Tang et al. (2018) found differences between droplet density and flight height. Spray performance at an operating height of 1.2 m was better than at other heights evaluated for lime crop. Studies by Lou et al. (2018) found interaction of flight height with droplet density for the cotton crop. Qin et al. (2018) also found that flight height influenced droplet density.

SPRAYING QUALITY USING UNMANNED AERIAL VEHICLE IN CITRUS
This influence leads to less droplet drift, but poor droplet deposition. This fact occurs mainly because the leaves, especially within the upper canopy, are influenced by the powerful downward wind, which makes it difficult for the droplets to adhere.
This investigation showed statistical differences for flight height below 4 m. Therefore, further studies should be conducted to evaluate the influence of flight height at lower levels, seeking expressive results in droplet density for lime crops.
Evaluating only the droplet density can determine the best quantity of droplets per area that would be appropriate for each application. However, the volumetric median diameter (VMD) is another crucial factor when defining which spray is more appropriate. Therefore, it should not be used alone as a requirement in the evaluation of agrochemical application (MEWES et al., 2013).
The F test values and significance, beyond the coefficient of variation (in percentage) of VMD in the analysis of variance are presented in Table  4. A significant difference for the spray nozzles rotation and flight height variables was found with no interaction between them. This result indicates that these variables act independently on the values of volumetric median diameter.
In Figure 2 we can observe the volumetric median diameter averages variation for the different spray nozzles rotation and flight heights. We verified a greater VMD for the rotation of 7,500 rpm when compared with the rotation of 10,500 rpm. This result indicates that the lower the spray nozzles rotation, the greater the VMD.
In Figure 2B we can note a significant difference between the flight heights of two and three meters. The flight height of four meters, in turn, did not differ statistically amongst them. At two meters, the highest VMD value was 301.08 µm.  2.08 ns CV = 13.46 % VMD = volumetric median diameter; ns = not significant at 5% probability level by F test; and *= significant at 5% probability level by F test Similar conclusions were reached for operating heights of 3.0 m and 4.0 m. In this sense, it can be inferred that smaller droplets can easily penetrate the canopy and be deposited in the innermost areas of the plant canopy, while larger droplets are deposited on the leaves in the higher and outer layers. However, small droplets tend to be more prone to drift than large droplets, so drift must be controlled (TANG et al., 2018).
Droplet diameter averages were greater than 100.00 μm, indicating environmentally safe spraying, since wider droplets present lower drift risks (ZAMPIROLI et al., 2019). Berna (2017) described that droplets less than 50 µm wide tend to evaporate before reaching the desired target. At times when wind speed is very low (< 3 km.h -1 ), sprayed droplets, especially the finest ones, can be suspended in the air, not reaching the target and being dispersed to outer areas (CONTIERO et al., 2018). Tang et al. (2018) found results that show the same trends in droplet coverage rate and droplet density. It is likely that the proper degree of downwash generated by the rotors will open the canopy and help the particles penetrate deeper into the plant canopy (XUE et al., 2014;XUE et al., 2016;CABRAL & VITÓRIA, 2022).
Each of the 43 different genotypes of lime crops presents different canopy architectures. Ongoing experiments are evaluating other genotypes, different ages of crop development and application rates best suited for each combination of these variables, as well as verifying the effectiveness of pesticides applied by unmanned aerial vehicles.

CONCLUSION
• Higher rotations of the centrifugal spray nozzles increased the coverage rate of the sprayed droplets for flight heights of 2.0 m and 3.0 m; • The droplet density was influenced by spray nozzle rotation and spraying showed high mean VMD.
• The results show that spray heights of 3.0 m and 4.0 m have a notable effect on the amount of deposition, but their influence on droplet size is negligible.