Field Evaluation of Different Unmanned Aerial Spraying Systems Applied to Control Panonychus citri in Mountainous Citrus Orchards

Abstract

In mountainous citrus orchards, the application of conventional ground sprayers for the control of citrus red mite (Panonychus citri) is often constrained by complex terrain and low operational efficiency. The Unmanned Aerial Spraying System (UASS), due to its low-altitude, low-volume, and high-maneuverability characteristics, has emerged as a promising alternative for pest management in such challenging environments. To evaluate the spray performance and field efficacy of different UASS types in controlling P. citri, five representative UASS models (JX25, DP, T1000, E-A2021, and T20), four mainstream pesticide formulations, and four novel tank-mix adjuvants were systematically assessed in a field experiment conducted in a typical hilly citrus orchard. The results showed that T20 delivered the best overall spray deposition, with upper canopy coverage reaching 10.63%, a deposition of 3.01 μg/cm², and the highest pesticide utilization (43.2%). E-A2021, equipped with a centrifugal nozzle, produced the finest droplets and highest droplet density (120.3–151.4 deposits/cm²), but its deposition and coverage were lowest due to drift. Nonetheless, it exhibited superior penetration (dIPR 72.3%, dDPR 73.5%), facilitating internal canopy coverage. T1000, operating at higher flight parameters, had the weakest deposition. Formulation type had a limited impact, with microemulsions (MEs) outperforming emulsifiable concentrates (ECs) and suspension concentrates (SCs). All adjuvants improved spray metrics, especially Yimanchu and Silwet, which enhanced pesticide utilization to 46.8% and 46.4% for E-A2021 and DP, respectively. Adjuvant use increased utilization by 4.6–11.9%, but also raised ground losses by 1.5–4.2%, except for Yimanchu, which reduced ground loss by 2.3%. In terms of control effect, the rapid efficacy (1–7 days after application, DAA) of UASS spraying was slightly lower than that of ground sprayers—electric spray gun (ESG), while its residual efficacy (14–25 DAA) was slightly higher. The addition of adjuvants improved both rapid and residual efficacy, making it comparable to or even better than ESG. E-A2021 with 5% abamectin·etoxazole ME (5A·E) and Yimanchu achieved 97.4% efficacy at 25 DAA. Among UASSs, T20 showed the rapid control, while E-A2021 outperformed JX25 and T1000 due to finer droplets effectively targeting P. citri. In residual control (14–25 DAA), JX25 with 45% bifenazate·etoxazole SC (45B·E) was most effective, followed by T20. 5A·E and 45B·E showed better residual efficacy than abamectin-based formulations, which declined more rapidly. Adjuvants significantly extended control duration, with Yimanchu performing best. This study demonstrates that with optimized spraying parameters, nozzle types, and adjuvants, UASSs can match or surpass ground spraying in P. citri control in hilly citrus orchards, providing valuable guidance for precision pesticide application in complex terrain.

1. Introduction

Citrus (Citrus reticulata Blanco) is an important economic forest fruit in southern China, and its planting scale has significant regional characteristics. According to agricultural remote sensing data, more than 90% of citrus planting areas are concentrated in the subtropical monsoon climate zone south of the Yangtze River, of which 60–70% of orchards are distributed in hilly and mountainous ecosystems with complex topography (average slope > 15°) [1]. This geomorphological feature leads to the dual challenges of canopy management: firstly, the microclimate heterogeneity caused by slope topography (local temperature gradient of 2.8 °C/100 m, relative humidity difference > 30%) aggravates the risk of spatial transmission of pests and diseases [2]; secondly, limited mechanical operating conditions (mechanized area < 40%) make traditional plant protection methods inefficient [3]. The canopy closure of hilly and mountainous orchards is generally > 0.7, and the leaf area index (LAI) exceeds 3, which will hinder spray penetration and reduce the plant protection effect [4]. The canopy density of the orchards used in this study generally exceeded 0.7. This relatively closed canopy structure provides an ideal habitat for pests such as citrus red mite (Panonychus citri (McGregor, 1916)) (Acari: Tetranychidae), Asian citrus psyllids (Diaphorina citri Kuwayama) and aphids (Aphidoidea), making pests serious in mountain citrus orchards and increasing the difficulty of management.

P. citri is one of the most destructive tetranychid mites affecting citrus production. By piercing and feeding on the surfaces of leaves and fruits, it causes visible damage to fruit appearance and leads to significant reductions in both yield and quality. Its spatial distribution within the canopy is highly dynamic: under hot and dry conditions, populations tend to concentrate on sun-exposed, well-ventilated leaves and fruit surfaces in the upper and outer canopy, whereas in more humid or shaded environments, they may shift to the lower inner canopy [5]. Therefore, the precise control of P. citri remains a critical component of integrated pest management in citrus orchards. Chemical control is still the most widely used and effective approach, and achieving uniform droplet deposition within the canopy is essential for ensuring control efficacy. We investigated the following three types of pesticides commonly used to control P. citri in citrus production areas in southern China. (1) Bifenazate, a selective acaricide from the hydrazine ester class, exhibits high efficacy against various life stages of P. citri, with particularly potent contact toxicity against adult P. citri [6]. It has moderate penetration capacity, provides rapid action within 24 h, and maintains residual activity for approximately 24–30 days [7]. (2) Etoxazole, a diphenyl oxazoline derivative, primarily disrupts the moulting process of eggs and immature P. citri [8]. It has relatively weak stomach and contact toxicity, the lowest penetration capability, and a slow onset of action, but it offers an extended residual period of 30–50 days [9]. (3) Abamectin, a macrocyclic lactone neurotoxin, exerts exceptional stomach and contact toxicity against adult and immature P. citri, though it is less effective against eggs. It has the strongest penetration capacity and the fastest knockdown effect, but its residual activity is shorter, lasting approximately 7–14 days [10]. The combination of bifenazate and etoxazole provides comprehensive coverage across the entire mite life cycle, balancing moderate quick action with prolonged residual control. Meanwhile, abamectin combined with etoxazole offers both rapid initial efficacy and an extended suppression of mite populations. All three have no significant systemic activity, but have excellent control effectiveness when applied using traditional ground-spray equipment.

However, traditional ground-based spraying techniques are highly reliant on manual labour, posing significant challenges in hilly and mountainous citrus orchards. Moreover, traditional spraying techniques are characterized by high labour intensity, low operational efficiency, poor pesticide utilization, and a tendency to cause environmental contamination, along with potential health risks to applicators [11]. Unmanned Aerial Spraying System (UASS) has become a pivotal technological tool in modern agricultural pest control due to their low-altitude hovering capability, low-volume spraying technology, operational flexibility, and high efficiency. By integrating precision spraying systems with multi-sensor fusion technology, UASSs demonstrate significant advantages in mountainous and hilly orchards [11,12]. Ground-imitating technology, in particular, enables UASSs to maintain stable flight heights over uneven terrain, ensuring uniform droplet deposition and enhancing pest-control effect [13]. Compared to traditional ground-based sprayers, UASSs offer greater adaptability, reduced pesticide usage, and lower environmental impact by minimizing pesticide drift and soil contamination. With their superior suitability for complex terrains and enhanced precision, UASS spraying technology is increasingly favoured by orchard farmers and has become a focal point of research in smart agriculture and precision plant protection.

Currently, research on UASS-based pesticide application in orchards has been steadily advancing, primarily focusing on three core areas. Firstly, optimizing operational parameters is essential to improving spraying precision. Key parameters such as operating height, operating speed, spray angle, and flow rate directly influence droplet deposition and canopy coverage [14]. There are many studies on operating parameters, and this article does not focus on them.

Secondly, the scientific combination of pesticides, formulations, and adjuvants has gained considerable attention, as optimized formulations can enhance control effect, reduce pesticide usage, and minimize environmental impact [15]. In UASS pesticide applications, different formulations exhibit distinct spray characteristics. Suspension concentrate (SC) generates larger droplet sizes, resulting in weaker penetration and permeability, making them more suitable for long-lasting pesticides. But emulsifiable concentrate (EC) produces smaller droplets with higher permeability and penetration, providing rapid action but with an increased risk of drift [16,17]. In contrast, microemulsion (ME) has the advantages of both, with enhanced droplet penetration and permeability, especially for UASS spraying in high-canopy-density orchards [18]. A novel class of nano-enabled pesticides has emerged in recent years, offering broad formulation versatility across nearly all conventional pesticide types. Nano-pesticides are characterized by high specific surface area, elevated bioactivity, and strong permeability, enabling effective pest control at reduced active ingredient dosages [19]. Such properties confer substantial potential advantages for application via UASSs. Adjuvants are known to significantly reduce spray-solution surface tension, thereby improving droplet adhesion and spreading on crop surfaces and enhancing deposition within the canopy. For example, Nongjianfei^® (as shown in Table 1 in Section 2.1.2) is a high-molecular-weight polymeric adjuvant primarily containing polyethylene oxide–polypropylene oxide (PEO–PPO) copolymers, which function as high-molecular-weight surfactants with anti-drift properties [16]. Similarly, Beidatong^® is an aerial-application-specific adjuvant based on methylated plant oil, which dissolves the epicuticular wax layer to enhance penetration and promotes droplet settling to reduce drift. Previous studies have shown that while the synergistic effect of Nongjianfei is slightly inferior to that of Beidatong, it is significantly greater than that of the conventional organosilicone adjuvant Y-20079 [20].

Thirdly, improvements to spraying systems, including nozzle type selection, droplet size regulation, and spray pattern adjustments, remain a critical research focus to ensure uniform distribution and effective canopy penetration [21]. To address the complexities of orchard terrains, many UASS manufacturers have developed specialized agricultural models equipped with advanced navigation and obstacle-avoidance systems. These UASSs often integrate high-precision real-time height sensing and ground-imitating technology, enabling automatic flight altitude adjustment based on canopy height and terrain variation. This capability ensures consistent spray coverage and enhances pest-control efficacy [22]. However, due to variations in tree species, pesticide properties, UASS models, and configurations, research conclusions are often fragmented, lacking universal applicability. Previous studies have demonstrated that sloped terrain can induce localized pressure differentials and airflow recirculation, which disrupt the vertical symmetry of UASS downwash airflow and lead to oblique diffusion patterns. These phenomena result in directional deviations in droplet deposition. Moreover, terrain undulations and canopy obstructions may cause droplet rebound and uplift, increasing drift risk and reducing effective deposition [23]. This study provides a preliminary investigation into these effects under mountainous orchard conditions. These issues are particularly pronounced when controlling P. citri, a minute species with strong reproductive capacity and a preference for inhabiting the inner canopy [8], posing significant challenges for UASS-based applications.

In this study, five distinct types of unmanned aerial spraying systems (UASSs) were selected for field trials to evaluate their spray performance and control efficacy against P. citri under mountainous orchard conditions. Four pesticide formulations and four tank-mix adjuvants specifically designed for aerial application were tested. The research objectives are as follows: (1) to evaluate the spraying characteristics of five UASSs, including droplet density, deposition, coverage, penetration, and pesticide utilization rate, to comprehensively assess their application performance; and (2) to assess the control effects of five UASSs applying different pesticide formulations against P. citri, comparing their performance with ground-based spraying equipment to clarify the advantages and limitations of UASS application in orchards.

2. Materials and Methods

2.1. Materials

2.1.1. Test Conditions

This experiment was conducted in September 2020 in Yongfu County, Guilin City, Guangxi Province, China (longitude: 110°05′49.06″, latitude: 25°04′34.43″). The test orchard is a typical mountainous citrus orchard (as shown in Figure 1) with an average slope of 33.5 ± 5.5° (vertical height: 71.0 ± 9.1 m, horizontal span: 110.6 ± 22.9 m). The soil type is weakly acidic yellow soil with moderate fertility and a relatively low organic matter content. The orchard was planted along contour lines with 4-year-old late-ripening Shatangju (Citrus reticulata) citrus trees trained into an open-centre canopy structure. The trees exhibited vigorous growth, with heights 1.2–1.7 m, canopy diameters 1.8–2.6 m, plant spacing 2.5–2.8 m, row spacing 1.6–2.7 m, and a planting density of approximately 900 trees/ha. A representative citrus tree from the orchard is shown in Figure S1 of the Supplementary Materials.

The LAI, measured using the SunScan Canopy Analysis System (Delta-T Devices Ltd., Burwell, Cambridge, UK), was 3.6 ± 0.3. According to existing studies, this corresponds to a canopy closure of 0.7–0.8 [24], which reduces internal ventilation, increases humidity, and significantly accelerates the reproduction and spread of pests such as P. citri. At the start of the experiment, the P. citri population density was high, with an average of 5–25 active mites per leaf (as shown in Figure 11a in Section 3.2).

Meteorological conditions remained stable during the experiment. A Kestrel 5500 digital meteorograph (Loftopia, LLC, Birmingham, MI, USA) was used to record environmental data during each treatment. The daily average temperature was 26.6 ± 1.8 °C, relative humidity was 61.4 ± 4.5% RH, and wind speed was 0.1–0.6 m/s. The high canopy closure and the favourable microclimate, characterized by high humidity and moderate temperatures, likely facilitated rapid P. citri proliferation, posing significant challenges for UASS spraying operations.

2.1.2. Experimental Chemicals

The chemicals used in this study included the pesticides, adjuvants, and tracers listed in Table 1 and shown in Figure S2 (Supplementary Materials). None of the four pesticides used in this study exhibit significant systemic activity, which presents a disadvantage for UASS-based application. Due to canopy shielding, droplet penetration into the inner canopy tends to diminish, limiting deposition in target areas. Evaluating the efficacy of UASS spraying against P. citri under such conditions was a central focus of this study. Among the tested formulations, 1.8% abamectin ME and 5% abamectin·etoxazole ME are classified as nano-enabled pesticides, which can effectively improve the targeted release and effective deposition of pesticides. The tracing agent Allura Red 85 is a commonly used water-soluble tracer, which is mixed with pesticides without changing the physical and chemical properties of the liquid [25]. Before spraying, pesticides, adjuvants, and the tracer were fully dissolved in distilled water to ensure the reliability of subsequent spray performance analysis.

The four adjuvants selected in this study were all specifically designed for aerial spraying tank-mix applications, as shown in Table 1. Among them, Chengji^® is an upgraded version of Beidatong, enhanced by the addition of organosilicone. Organosilicone exhibits strong spreading ability and facilitates pesticide penetration into plant stomata. Therefore, Chengji exhibits superior wettability and permeability, making it more suitable for aerial spraying in orchard environments. Silwet 510^® is a modified trisiloxane based on hydroxyl silicone oil, which is a new type of organosilicon adjuvant. Unlike conventional organosilicon adjuvants, Silwet 510 exhibits lower penetration capacity but stronger adhesion with minimal foaming, thereby improving spray solution spreading and rainfastness. Yimanchu^® is a functional adjuvant formulated as a blend of organosilicon surfactants, anionic surfactants, and plant essential oils. When applied alone, Yimanchu exerts rapid contact toxicity against mites by physically disrupting their respiratory systems. When mixed with pesticides, it significantly enhances pesticide penetration, spreading, and wetting properties. The field trial further evaluated the synergistic effects of these four adjuvants on aerial spray performance.

Table 1. Details of aviation spray pesticides and adjuvants.

Reagent Type	Name and Specifications	Manufacturer	Application Dosage	Active Ingredient Content (g/ha)
pesticides	45% Bifenazate·Etoxazole SC (45B·E)	Guilin Jiqi Group Co., Ltd., Guilin, Guangxi, China	1500 mL/ha	675
	5% Abamectin EC (5AVM)	Hebei Veyong Bio-chemical Co., Ltd., Shijiazhuang, Hebei, China	1200 mL/ha	60
	1.8% Abamectin ME (1.8AVM)	Nanjing Sense Biotechnology Co., Ltd., Nanjing, Jiangsu, China	1005 g/ha	18.09
	5% Abamectin·Etoxazole ME (5A·E)		1500 g/ha	75
adjuvants	Nongjianfei	Guilin Jiqi Group Co., Ltd., Guilin, China	120 mL/ha	/
	Silwet 510	Momentiveperformance materials (Shanghai) Co., Ltd., Shanghai, China	600 mL/ha	/
	Chengji	Hebei Mingshun Agricultural Technology Co., Ltd., Shijiazhuang, Hebei, China	600 mL/ha	/
	Yimanchu	Chongqing Lingshi Agricultural Technology Co., Ltd., Chongqing, China	1500 mL/ha	/
tracer agent	Allura Red 85	Zhejiang Dragoni Colour Technology Co., Ltd., Longgang, Zhejiang, China	450 g/ha	/

Table 1. Details of aviation spray pesticides and adjuvants.

Reagent Type	Name and Specifications	Manufacturer	Application Dosage	Active Ingredient Content (g/ha)
pesticides	45% Bifenazate·Etoxazole SC (45B·E)	Guilin Jiqi Group Co., Ltd., Guilin, Guangxi, China	1500 mL/ha	675
	5% Abamectin EC (5AVM)	Hebei Veyong Bio-chemical Co., Ltd., Shijiazhuang, Hebei, China	1200 mL/ha	60
	1.8% Abamectin ME (1.8AVM)	Nanjing Sense Biotechnology Co., Ltd., Nanjing, Jiangsu, China	1005 g/ha	18.09
	5% Abamectin·Etoxazole ME (5A·E)		1500 g/ha	75
adjuvants	Nongjianfei	Guilin Jiqi Group Co., Ltd., Guilin, China	120 mL/ha	/
	Silwet 510	Momentiveperformance materials (Shanghai) Co., Ltd., Shanghai, China	600 mL/ha	/
	Chengji	Hebei Mingshun Agricultural Technology Co., Ltd., Shijiazhuang, Hebei, China	600 mL/ha	/
	Yimanchu	Chongqing Lingshi Agricultural Technology Co., Ltd., Chongqing, China	1500 mL/ha	/
tracer agent	Allura Red 85	Zhejiang Dragoni Colour Technology Co., Ltd., Longgang, Zhejiang, China	450 g/ha	/

2.1.3. Spray Equipment

Five commercial UASSs were used: Jinxing 25 (JX25, Wuxi Hanhe Aviation Technology Co., Ltd., Wuxi, Jiangsu, China), E-A2021 (Eavision Technology Co., Ltd., Suzhou, China), Agras T20 (T20, DJI Technology Co., Ltd., Shenzhen, China), Free Eagle DP (DP, Anyang Quanfeng Aviation Plant Protection Technology Co., Ltd., Anyang, Henan, China), and Global Hawk T1000 (T1000, Anyang Quanfeng Biological Technology Co., Ltd., Anyang, Henan, China), as shown in Figure 2 and Figures S3 and S4 (Supplementary Materials). Among them, JX25 and E-A2020 are electric four-rotor model UASSs, T20 and DP are electric six-rotor models, and T1000 is a petrol-powered single-rotor model. Specific parameters are shown in

Table 2. Main parameters of the five UASS models.

Parameters	JX25	E-A2021	T20	DP	T1000
Dimensions (mm)	1235 × 1235 × 647	1430 × 1170 × 510	1795 × 1510 × 732	2485 × 2255 × 850	2130 × 700 × 670
Total weight (kg)	23.5	25.25	21.1	42	30
Tank volume (L)	22	20	20	18	12
Nozzle type	Teejet 110025vk	CCMS	Teejet SX11001VS	Teejet XR11001VS	Teejet XR11001VS
Number of nozzles	4	2	8	8	5
Droplet size (μm)	120–200	20–250	130–250	130–250	130–250
Flow rate (L/min)	3.2–7.4	0.4–3.5	3.6	4–7	1.44–1.86
Effective swath width (m)	6–7	4–5	4–7	5.5–6.5	4–6

. The E-A2020 is equipped with centrifugal nozzles, producing a narrow droplet spectrum with small and uniform droplets [26]. In contrast, the remaining UASSs use hydraulic atomizing nozzles, which have a simpler structure and generate larger droplet sizes with higher flow rates [12]. Each UASS was equipped with a GNSS + RTK dual-redundancy positioning system, advanced sensor technology, and an autonomous flight control system. These features enabled fully autonomous task planning, terrain-following flight, and obstacle avoidance, making them ideal for orchard operations. This mode is commonly referred to as the “Orchard Operation Mode” [25,27]. Due to the steep slope at the test site and surrounding obstacles such as mountains and tall trees, all five UASSs employed the “A–B” waypoint planning method, using a semi-autonomous terrain-following mode [28] for the operations. In recent years, unmanned aerial spraying systems (UASS) have undergone rapid technological advancements, particularly in payload capacity, terrain-adaptive spraying, and atomization techniques. Notably, DJI’s T50 model integrates dual atomization nozzles combining hydraulic and centrifugal mechanisms, enabling more precise real-time terrain-following applications [29]. Nevertheless, earlier-generation models such as the T16 and T20, which support orchard-mode operations, continue to be widely adopted by growers due to their cost-effectiveness and operational stability.

For comparison, a stretcher-mounted electric spray gun (ESG, Shanghai Yueji Motor Co., Ltd., Shanghai, China) was used in this experiment as a comparison with the UASSs (Figure 2f). The external dimensions of the sprayer are 0.72 × 0.37 × 0.40 m, and it is equipped with a 3WZB-60 three-cylinder plunger pump with a theoretical flow rate of 46–60 L/min. The spray tank has a capacity of 400 L, and the operating pressure ranges from 1.0 to 4.0 MPa.

2.2. Experiment Design and Methods

2.2.1. Experiment Treatment Planning

The experiment included a total of 16 treatment combinations, each with two replications. Detailed settings are presented in

Table 3. Experiment treatment design.

Treatment	Spray Equipment	Testing Pesticides	Effective Dose (g a.i./ha)	Adjuvants (mL)	Flight Speed (m/s)	Flight Height (m)
1	JX25	45B·E	675	/	3	1.5
2				Nongjianfei 8
3	DP	5AVM	60	/	4	3
4				Silwet 40
5	T1000	1.8AVM	18.09	/	5	4
6				Chengji 40
7	E-A2021	5A·E	75	/	3	3
8				Yimanchu 100
9	T20	45B·E	675	/	3	3
10		5AVM	60	/
11		1.8AVM	18.09	/
12		5A·E	75	/
13	ESG	45B·E	675	/	/	/
14		5AVM	60	/	/	/
15		1.8AVM	18.09	/	/	/
16		5A·E	75	/	/	/

. Each UASS applied a spray volume of 60 L/ha, and other operating parameters were based on the best fruit-tree operating parameters provided by UASS manufacturers. The spraying volume of the ground-spraying equipment was 3000 L/ha, and the spraying method was consistent with the local agricultural production practice.

2.2.2. Spray Performance Test Sampling

Firstly, the orchard was divided into 32 test plots, each consisting of two adjacent rows of citrus trees along the same contour line, with an area of approximately 700 m² per plot. Due to the limited total orchard area and low wind speed during the experiment, a single row of trees served as a buffer zone between test plots to prevent spray drift from interfering with the results. The UASS flight paths were directly over the rows of trees. In each plot, three non-adjacent trees were selected for droplet deposition sampling. All the sampled fruit trees grew similarly and the tree shapes were similar. Two trees were positioned approximately 30 m from the start and end points of the UASS flight path, and one was located at the centre of the plot. The distance between any two sampled trees was no less than 30 m. As illustrated in Figure 3, the canopy of each tree was divided into three sampling layers: the upper layer (10 cm from the top canopy at the canopy edge), the inner layer (midway between the upper and lower layers, approximately 40 cm from the trunk), and the lower layer (50 cm above ground at the canopy edge). In each sampling layer, one healthy and well-formed leaf was selected in four directions—front, back, left, and right (with the flight direction as the front) of the tree. A droplet test card (DC, coated paper, 3 cm × 7 cm, Nanjing Oracle Technology Co., Ltd., Nanjing, Jiangsu, China) was stapled to the sunward side of each leaf. It is important to note that the length of the droplet test card was similar to the leaf length and was placed parallel to the leaf surface. Due to the high humidity in the orchard during the experiment, which caused water-sensitive paper to blur easily, coated paper was used to collect droplets as it dries more quickly. In the canopy’s ground projection area, a Mylar card (MC, 5 cm × 8 cm) was pinned in each of the four directions (front, back, left, and right). Additionally, five MCs were placed on the ground between adjacent trees in the forward direction (parallel to the flight path) and to the right (perpendicular to the flight path) to collect runoff and assess spray loss.

Due to the steep slope of the orchard, all UASSs adopt a semi-automatic terrain-following flight operation mode for safety. That is, the starting point of each test area is marked as ‘A’ and ‘B’ points on the UASS’s remote control, and the flight path followed the contour lines, flying back and forth directly above each row of citrus trees. During each test treatment, a certain amount of Allura Red was added to the solution as an indicator, and 10 L of solution was added to each UASS for spraying. After the liquid was dried, five citrus leaves of similar size were picked near the position where the DC was arranged, placed in different 5# self-sealing bags, and the DC was collected and stored in an envelope. The MCs placed under the citrus trees were recycled, and each piece was placed in a different 5# self-sealing bag. The samples were properly labelled and placed in black self-sealing bags before being taken back to the laboratory. Additionally, in the control area, the ground-based spraying equipment used was a ESG operated by local farmers, with the nozzle positioned approximately 30 cm from the plants, ensuring uniform spraying across every corner of the canopy. This study aimed to compare the spray characteristics of different UASSs, focusing only on the control efficacy in comparison with ground-based spraying equipment; therefore, no droplet collectors were placed in the control area.

The P. citri population was surveyed 1 day before spraying, and the number of live mites was recorded on days 1, 3, 7, 14, and 25 after spraying. Following the pesticide field efficacy test standard GB/T 17980.11-2000 [30], five non-adjacent citrus trees were randomly selected from each plot for investigation. On each tree, young shoots were marked at five locations—east, west, south, north, and centre. Using a handheld magnifying glass, live mites were counted on both sides of the leaves. Five leaves were examined per shoot, resulting in a total count of live mites on 25 leaves per tree. Additionally, the effects of pesticide application on the growth of citrus fruits, leaves, and young shoots were observed.

2.2.3. Sample Processing and Statistical Analysis

At the end of the experiment, the samples were immediately stored in a refrigerator and processed within one week. The DCs were scanned into 600 dpi grayscale images using a Microtek ScanMaker i2000 scanner (Shanghai Microtek Technology Co., Ltd., Shanghai, China). The grayscale images were then processed using the DepositScan^TM software (USDA-ARS, Wooster, OH, USA) to obtain the parameters for coverage and droplet density [31]. In this experiment, the pesticide application rates varied across different treatments. Therefore, the spray-quality parameters of the tracer, such as droplet coverage, droplet density, and droplet deposition, were used to represent the relevant parameters of the pesticide solution. The coefficient of variation (CV, %) was used to quantify droplet deposition uniformity, where a smaller CV indicates better uniformity of droplet deposition, as shown in Equation (1).

$$CV={\displaystyle \frac{s}{\overline{x}}}, S=\sqrt{{\displaystyle \frac{{\Sigma }_{i=1}^n{\left(x_i-\overline{x}\right)}^2}{n-1}}}$$

(1)

where S is the standard deviation; x_i refers to the spray-quality parameter at sampling point i; $\overline{x}$ denotes the average value of spray-quality parameter at all sampling points; n is the total number of sampling points.

The test was conducted under hot and dry conditions, which favoured the aggregation of P. citri on the upper and outer canopy. This created favourable conditions for UASS spraying. However, a small number of mites may still inhabit the lower inner canopy. Thus, improving droplet penetration into the middle and lower canopy is essential for effective pest control [11]. In this study, the ratio of droplet coverage or deposition in the inner and lower canopy layers to that in the upper layer was used as an indicator of droplet penetration ratio, namely the Inward Penetration Ratio (IPR) and Downward Penetration Ratio (DPR), to evaluate the ability of droplets to diffuse into the inner and lower canopy [32]. Droplet penetration ratio calculated based on coverage was defined as the coverage-based Inward Penetration Ratio (cIPR) and coverage-based Downward Penetration Ratio (cDPR), whereas droplet penetration ratio based on deposition amount was defined as the deposition-based Inward Penetration Ratio (dIPR) and deposition-based Downward Penetration Ratio (dDPR). These parameters provide a quantitative assessment of droplet transport within the canopy, offering insights for optimizing UASS spraying strategies to enhance pest-control efficacy.

Droplet deposition was measured according to ISO 24253-1 standards [33]. In each self-sealing bag containing the leaf samples, 20 mL of ultrapure water was added and the samples were thoroughly shaken to wash the leaves, ensuring that the was fully dissolved in the water. The absorbance of the eluate was then measured at a wavelength of 514 nm using an SMP500 MD microplate reader (MD Electronics, Inc., Poulsbo, WA, USA). The concentration of Allura Red in the eluent was calculated based on the standard curve equation, Y = 0.0277X + 0.0556, R² = 0.999, meeting the requirements for quantitative analysis. Subsequently, the deposition (μg/cm²) of Allura Red on the citrus canopy was determined using the volume of the eluent and the DC area, as described by Equation (2). In accordance with the NY/T 3630.1-2020 standard [34], the total deposition in each treatment area was then calculated based on the relationship between LAI and canopy projection area, enabling the determination of pesticide utilization rate (%) as described by Equation (3).

$${\beta }_{dep}={\displaystyle \frac{\left({\rho }_{smpl}-{\rho }_{blk}\right)\times F_{cal}\times V_{dil}}{A_{col}}}$$

(2)

$${\beta }_{dep\%}={\displaystyle \frac{{\overline{\beta }}_{dep}\times S_{grd}\times LAI\times \rho \times {10}^4}{M\times {10}^6}}\times 100\%$$

(3)

where β_dep is droplet deposition, μg/cm²; ρ_smpl is the absorbance value of the sample eluate; ρ_blk is the absorbance value of the blank control sample; F_cal is the value of the slope of the standard curve; V_dil is the volume of eluate, mL; A_col is the total area of sampled leaves, cm²; β_dep% is the pesticide utilization rate, %; ${\overline{\beta }}_{\mathit{dep}}$ is the average canopy droplet deposition, μg/cm²; S_grd is the canopy ground projection area, m²; LAI is the leaf area index; ρ is the planting density, plant/ha; and M is the application rate of Allura Red per unit area, g/ha.

A total of 5 mL distilled water was added to the self-sealing bags containing the MCs, and the mixture was thoroughly shaken to ensure that the Allura Red completely dissolved. The absorbance of the eluent was then measured at 514 nm using a microplate reader. The mass concentration of Allura Red in the eluent was calculated based on the standard curve, and the deposition of Allura Red on each MC was determined. This information was then used to calculate the ground deposition of the spray solution relative to the canopy projection area and the space between the trees. The ground deposition as a proportion of the total application of Allura Red represents the ground loss rate (%), as shown in Equation (4).

$${\eta }_{dep\%}={\displaystyle \frac{({\beta }_{dep1}\times S_{grd}\times \rho +{\beta }_{dep2}\times S_{blk})\times {10}^4}{M\times {10}^6}}\times 100\%$$

(4)

where η_dep% is the ground loss rate, %; β_dep1 is the amount of droplet deposition in the ground projection of the canopy, μg/cm²; β_dep2 is the amount of droplet deposition in the blanking zone between the trees, μg/cm²; and S_blk is the area of the space zone between the trees.

According to the GB/T 17980.11-2000 standard [30], the population reduction rate of active P. citri in each area before and after spraying was calculated, which in turn allowed for the determination of the corrected control efficacy (%). The formula is as follows:

$$R(\%)={\displaystyle \frac{N_0-N_1}{N_0}}\times 100, E(\%)={\displaystyle \frac{R_{treat}-R_{CK}}{100-R_{CK}}}\times 100$$

(5)

where R is pest population reduction rate, %; N₀ is the number of live insects before spraying; N₁ is the number of live insects after spraying; E is the corrected control efficacy, %; R_treat is the pest population reduction rate in the spraying zone, and R_CK is the pest population reduction rate in the controlled zone.

Excel 2021 software (Microsoft Corp., Redmond, WA, USA) was used for data analysis. Given the complexity of the experimental design, the combination of “sprayer + pesticide + adjuvant” was treated as an independent treatment unit. A one-way analysis of variance (ANOVA) under a completely randomized design was adopted as the analytical approach for this study. As numerous previous studies have investigated the individual effects of different sprayers, pesticide formulations, and adjuvants, interaction effects among these factors were not considered in the present analysis. All statistical analyses were performed using DPS software (v7.05; Refine Information Technology Co., Ltd., Hangzhou, China). Additionally, Duncan’s multiple range test was used for significance testing, with the significance level set at p < 0.05. The figures were constructed using Origin 2021 (OriginLab Co., Ltd., Northampton, MA, USA) software package.

3. Results and Discussion

3.1. Results of Spraying Performance

3.1.1. Distribution of Canopy Droplet Density

Due to the influence of impact force, surface tension, and drying effects, droplet spreading and coalescence upon contact with DC result in measured droplet sizes being larger than the manufacturer’s specifications (Table 2). Additionally, edge-detection errors in the DepositScan software further contribute to the overestimation of DV₅₀ [31], as shown in Figure S5 (Supplementary Materials). Droplet density and coverage offer more meaningful insights for spray characterization, with density distributions under different treatment conditions presented in Figure 4. The droplet density distribution under different treatment conditions is shown in Figure 4. Generally, due to canopy obstruction, airflow turbulence, and operational constraints, UASS spraying exhibits a gradual decrease in droplet density, coverage, and deposition from the upper to the lower canopy and from the outer to the inner regions [14]. However, the E-A2021 UASS demonstrated significantly higher droplet density compared to the other UASSs (p < 0.001), with a greater number of droplets reaching the inner and lower canopy layers than the upper layer, resulting in an increasing trend in droplet density from top to bottom. Without the addition of adjuvants, the droplet densities in the upper, middle, and lower canopy layers for E-A2021 were 120.3 deposits/cm², 133.1 deposits/cm², and 151.4 deposits/cm², respectively. Because E-A2021 is equipped with a centrifugal nozzle (CCMS), the fine and dense droplets produced by the spray are difficult to attach to the leaves in the upper part of the canopy due to air turbulence, but more small droplets diffuse into the inner and lower canopy [26]. In contrast, the other four UASSs utilized pressure-swirl nozzles, and among them, T20 exhibited the highest droplet density without adjuvants. When spraying 5A·E, the droplet density was the highest, with 48.7 deposits/cm², 46.0 deposits/cm², and 45.7 deposits/cm² in the upper, middle, and lower canopy layers, respectively. However, spraying 45B·E resulted in the lowest droplet density, though the differences were not statistically significant. The droplet densities in JX25, DP, and T1000 were all lower than those in T20, with differences ranging from 0.5 to 10.8, 8.4–12.4, and 8.6–11.9 deposits/cm² across canopy layers, respectively.

After the addition of adjuvants, droplet density decreased by 5.2–6.9 deposits/cm², 3.6–9.2 deposits/cm², and 15.0–20.8 deposits/cm² when JX25, T1000, and E-A2021 applied pesticide solutions containing Nongjianfei, Chengji, and Yimanchu adjuvants, respectively. Notably, Yimanchu significantly reduced droplet density in both the upper and inner canopy layers of citrus trees (p < 0.001). In general, adjuvants reduce droplet surface tension, leading to a decrease in initial droplet size, which theoretically should increase droplet density. However, excessive spreading promotes rapid coalescence, reducing the number of independent droplets per unit area [35]. Interestingly, after DP sprayed the formulation containing Silwet 510, droplet density increased by 1.3–6.5 deposits/cm² (4–18%). This phenomenon is likely due to Silwet 510’s lower permeability compared to conventional organosilicon adjuvants, which slows the spreading rate on the target surface, preventing overexpansion and excessive droplet coalescence. Additionally, its high adhesion properties reduce droplet rebound and coalescence upon impact, ensuring that more independent droplets remain attached to the target surface, thereby enhancing overall droplet density. This characteristic makes it more suitable for nozzles with larger spray droplet sizes. At present, there is no research on the application of Silwet 510 adjuvant in pesticide spray, but there are many studies on other adjuvants of Silwet series. Among them, organosilicon adjuvants such as Silwet L-77 and Silwet DRS-60 have demonstrated the ability to enhance droplet density even under conditions of relatively low penetration, with this effect being particularly pronounced at lower concentrations, indicating a clear dose-dependent response [36]. Notably, a study by Hołownicki et al. [37] revealed that when Silwet L-77 was applied at a concentration of 0.5 mL·L⁻¹, droplet density increased by approximately 10% to 15%. Although systematic data on Silwet 510 remain unavailable, its observed influence on droplet density appears consistent with these findings, suggesting it may exert a similar surfactant-mediated regulatory effect on spray dynamics.

3.1.2. Distribution of Canopy Droplet Coverage and Penetration

As shown in Figure 5, the distribution of droplet coverage within the citrus canopy was assessed for five UASSs, all operating under the same spray volume of 60 L·ha⁻¹. The results, derived from an image analysis of DC, revealed a clear stratification pattern. Specifically, droplet coverage in the upper, inner, and lower canopy layers ranged from 8.96 to 12.68%, 5.56–8.91%, and 4.64–8.69%, respectively. In all treatments, coverage in the upper canopy was substantially higher than in the inner and lower layers, with the inner layer exhibiting slightly greater coverage than the lower one. When spraying without adjuvants, the UASS models demonstrated significant variation in canopy coverage. Unlike the trend observed for droplet density, the T20 produced the highest droplet coverage across all canopy layers. In comparison, the JX25, DP, T1000, and E-A2021 showed reductions in coverage relative to the T20 by approximately 0.97–1.18%, 1.01–1.31%, 1.04–1.38%, and 1.42–2.44%, respectively. Among them, the E-A2021 exhibited the lowest performance, with canopy coverage values of 8.19%, 5.57%, and 5.43% in the upper, inner, and lower layers, respectively. Notably, its inner canopy coverage was significantly lower than that of the T20 (F = 3.98, p < 0.001).

Under consistent flight parameters and spray volume, the T20 sprayer was used to apply four different pesticide formulations without the use of adjuvants. Among these, the application of 5A·E resulted in the highest droplet coverage, with coverage values of 10.63%, 7.51%, and 6.85% in the upper, inner, and lower canopy layers, respectively. The treatments with 5AVM and 1.8AVM followed, while 45B·E exhibited the lowest coverage values of 9.94%, 6.73%, and 5.63%, respectively. However, the differences among treatments were not statistically significant. Following the addition of adjuvants, droplet coverage increased across all canopy layers. Since the influence of pesticide formulation type on droplet coverage and deposition was minimal, the comparative enhancement effects of the adjuvants could be reasonably evaluated [16]. In the upper canopy layer, Nongjianfei, Silwet, Chengji, and Yimanchu increased droplet coverage by 0.90%, 3.10%, 2.42%, and 3.23%, respectively. In the inner canopy, the increases were 0.61%, 2.68%, 1.98%, and 3.09%, while in the lower canopy, the corresponding increases were 0.53%, 2.38%, 2.10%, and 3.26%.

The evaluation of droplet penetration based on droplet coverage is shown in Figure 6. When using the T20 sprayer, the cIPR values for 5A·E and 5AVM were 70.6% and 70.4%, and the corresponding cDPR values were 64.4% and 61.6%, respectively. These values were higher than those observed for 1.8AVM and 45B·E, whose cIPR values were 68.6% and 67.7%, and cDPR values were 60.5% and 56.7%, respectively. For the JX25, DP, T1000, and E-A2021 UASSs without adjuvant addition, the cIPR values were 62.0%, 65.1%, 63.0%, and 68.1%, while the corresponding cDPR values were 51.8%, 54.5%, 57.7%, and 66.3%, respectively. Among the five UASSs, T20 achieved the highest cIPR, while E-A2021 recorded the highest cDPR. After adjuvant addition, cIPR values increased by 0.5%, 5.2%, 4.1%, and 7.8%, and cDPR values increased by 0.6%, 5.4%, 6.2%, and 9.8%, respectively. These results further support the conclusions drawn in the previous section.

3.1.3. Distribution of Canopy Droplet Deposition and Penetration

The five UASSs were operated at a spray volume of 60 L·ha⁻¹, with Allura Red applied at a dosage of 450 g·ha⁻¹ as a tracer. The deposition of Allura Red was quantified using ultraviolet spectrophotometry. The theoretical maximum deposition of the tracer within the treated area was no more than 4.50 μg·cm⁻². The spatial distribution of tracer deposition was used to characterize spray-liquid deposition across the canopy, as illustrated in Figure 7. Overall, droplet deposition exhibited a vertical stratification pattern, with values of 1.87–3.42 μg·cm⁻², 1.35–2.49 μg·cm⁻², and 1.36–2.14 μg·cm⁻² in the upper, inner, and lower canopy layers, respectively. The distribution of deposition closely mirrored that of droplet coverage, with significantly higher values observed in the upper canopy compared to the inner and lower layers, and slightly greater deposition in the inner layer relative to the lower. Among the five UASSs tested without adjuvant application, the T20 achieved the highest droplet deposition. In comparison to T20, the droplet deposition values of JX25, DP, T1000, and E-A2021 were reduced by approximately 0.16–0.23 μg·cm⁻², 0.20–0.34 μg·cm⁻², 0.24–0.48 μg·cm⁻², and 0.61–1.14 μg·cm⁻², respectively. E-A2021 exhibited the lowest deposition overall, with values of 1.87 μg·cm⁻², 1.35 μg·cm⁻², and 1.38 μg·cm⁻² in the upper, inner, and lower canopy layers, respectively. In particular, its deposition in the upper and inner layers was significantly lower than that of T20 (p < 0.001). These results indicate that, despite producing a much higher droplet density than the other UASSs, the E-A2021, equipped with a centrifugal nozzle, resulted in the lowest droplet coverage and deposition. This paradox indicates that due to the fine and dense droplets formed by centrifugal nozzles, the droplets are subject to excessive drift or evaporative loss, which reduces the effective deposition of the droplets [26]. Both DP and T1000 were equipped with the same fan-shaped nozzles as the T20 but operated at higher spray parameters (DP: speed 4 m·s⁻¹, height 3 m; T1000: speed 5 m·s⁻¹, height 4 m), likely reducing the number of droplets reaching the target and contributing to lower coverage and deposition. The JX25, fitted with a cone-type nozzle and operating at lower parameters (speed 3 m·s⁻¹, height 1.5 m), also exhibited lower droplet density, coverage, and deposition compared to T20. This may be due to differences in the spraying system and the use of a SC formulation. The JX25, equipped with a cone nozzle and operated at relatively low parameters (flight speed: 3 m·s⁻¹; height: 1.5 m), exhibited lower droplet deposition than the T20, which may be attributed to differences in the spray system and the use of SC formulations. In addition, sloped terrain and canopy obstructions can significantly disturb UASS downwash airflow, causing oblique dispersion, recirculation, or the uplift of droplets, thereby reducing targeted deposition efficiency [23]. As a result, even with similar nozzle configurations, the interaction between operational parameters and terrain conditions may lead to considerable variation in spray performance across UASS models. This effect is particularly pronounced for E-A2021, which emits fine and dense droplets, and for high-speed systems like DP and T1000. The combined influence of nozzle type, flight parameters, and orchard topography significantly influence the overall efficacy of UASS-based spraying, which is consistent with findings reported by Baio [26] and Zeeshan [16]. However, Liu et al. [38] observed that under higher operational parameters (speed: 4 m·s⁻¹, height: 3–4 m), cone-type nozzles outperformed fan-shaped nozzles in terms of droplet coverage and deposition. Therefore, for cone-nozzle UASSs such as the JX25, increasing flight height and speed may improve deposition, whereas the opposite adjustment may be more suitable for the DP and T1000, particularly in steep orchard terrains. Based on comprehensive results for both deposition and coverage, the overall spray quality of the five UASSs was ranked as follows: T20 > JX25 > DP > T1000 > E-A2021.

Under identical operating parameters, the T20 achieved the highest droplet deposition when spraying 5A·E without adjuvant application, with deposition levels in the upper, middle, and lower canopy layers reaching 3.01 μg/cm², 2.18 μg/cm², and 1.98 μg/cm², respectively. The deposition results for 5AVM and 1.8AVM followed closely, whereas the lowest deposition was observed with 45B·E, at 2.74 μg/cm² (upper), 1.69 μg/cm² (middle), and 1.98 μg/cm² (lower), with no statistically significant differences among treatments. Overall, ME and EC demonstrated superior deposition performance compared to SC in UASS-based spraying. However, the influence of pesticide formulation type on droplet coverage and deposition was less pronounced than that of operational parameters and adjuvants, which is consistent with the research of Zeeshan et al. [16]. In a vineyard study, Anken et al. [29] reported that newer UASS models such as the T50, equipped with rotary atomizers and operated at a height of 2.5 m and speed of 3 m/s, exhibited improved droplet distribution uniformity compared to earlier models like the T20. However, due to differences in experimental conditions and crop type, their findings are not directly comparable to those of the present study. Notably, their results also revealed insufficient droplet coverage within the inner canopy under high disease pressure (e.g., during peak periods of grapevine mildew). Furthermore, the study showed that the droplet deposition from T50 was only 14.1–27.8% of that achieved by backpack sprayers, indicating that UASS alone may not yet fully replace ground-based equipment.

Consistent with the trend observed for droplet coverage, the addition of adjuvants resulted in an overall increase in droplet deposition. In the upper canopy, Nongjianfei, Silwet, Chengji, and Yimanchu increased deposition by 0.44 μg/cm², 0.68 μg/cm², 0.66 μg/cm², and 0.71 μg/cm², respectively. In the inner canopy, the corresponding increases were 0.30 μg/cm², 0.61 μg/cm², 0.50 μg/cm², and 0.68 μg/cm². For the lower canopy, the deposition was enhanced by 0.29 μg/cm², 0.58 μg/cm², 0.57 μg/cm², and 0.73 μg/cm², respectively. Based on both droplet coverage and deposition results, the enhancement effects of the four adjuvants followed the order of Yimanchu > Silwet > Chengji > Nongjianfei. Notably, when the E-A2021 was operated with the Yimanchu adjuvant (speed: 3 m·s⁻¹; height: 3 m; spray volume: 60 L·ha⁻¹), the droplet coverage in the upper, inner, and lower canopy layers increased by 39.5%, 55.4%, and 60.0%, respectively, indicating a substantial improvement in droplet deposition within the canopy interior. This appears contradictory to the droplet density reduction described in Section 3.1.1. As the use of adjuvants decreases droplet size, increasing the initial droplet count. However, under the downward airflow of UASS, these finer droplets are more prone to collision and coalescence, effectively reducing drift and evaporation losses [35]. In a study by Guo et al. [14], when E-A2021 was operated under lower parameters (speed: 2.8 m·s⁻¹; height: 2.5 m; spray volume: 60 L·ha⁻¹), the addition of a non-silicone blended adjuvant (Lifei D) increased droplet coverage from 1.98% to 3.00%, representing an increase of approximately 52%. In comparison, Yimanchu exhibited superior efficacy in enhancing droplet penetration and deposition within the canopy.

As shown in Figure 8, the evaluation of droplet penetration based on deposition (dIPR and dDPR) revealed that T20 spraying with 5A·E and 5AVM achieved higher dIPR values (72.5% and 71.2%) and dDPR values (65.8% and 63.7%) than spraying with 1.8AVM and 45B·E (dIPR: 67.6% and 61.6%; dDPR: 63.1% and 58.1%). This phenomenon may be attributed to differences in the surface tension characteristics of the formulations. Microemulsions and emulsifiable concentrates with relatively low dynamic surface tension (<35 mN/m) may promote droplet wetting and redistribution within the canopy, whereas SCs with high zeta potential (>30 mV) tend to induce droplet coalescence, increase droplet size, and ultimately reduce canopy penetration efficiency [39]. Under non-adjuvant conditions, the dIPR values for JX25, DP, T1000, and E-A2021 were 58.6%, 68.6%, 66.5%, and 72.3%, respectively, while the corresponding dDPR values were 52.6%, 56.2%, 65.7%, and 73.5%. It is noteworthy that although E-A2021 exhibited dIPR values comparable to those of T20, its dDPR was significantly higher, likely due to the fine droplets generated by centrifugal nozzles, which more readily penetrate the vertical canopy structure under gravity [26]. Among the five UASSs, T20 and E-A2021 showed the highest dIPR values, with E-A2021 achieving the highest dDPR. Following adjuvant addition, dIPR values increased by 1.5%, 4.1%, 2.1%, and 6.4% for JX25, DP, T1000, and E-A2021, respectively; while dDPR values increased by 2.0%, 5.7%, 4.4%, and 7.9%. The most pronounced improvements were observed in E-A2021, indicating that adjuvants had a synergistic regulatory effect on its fine droplet system—by increasing droplet size and enhancing resistance to evaporation, adjuvants may help to balance the trade-off between droplet deposition and drift.

3.1.4. Canopy Pesticide Utilization Rate and Ground Loss Rate

As shown in the bar chart in Figure 9, pesticide utilization rate was evaluated using the deposition utilization rate of the tracer dye Allura Red. The results revealed that the pesticide utilization rate across the five UASSs ranged from 32.5% to 46.8%, significantly lower than that of conventional ground-spraying methods, which typically achieve 60–80% efficiency [14]. When sprayed without adjuvants, the T20 had the highest pesticide utilization rate, and the T20 was slightly higher when spraying 5A·E and 5AVM (43.2% and 42.0%, respectively) than when spraying 1.8AVM and 45B·E (41.2% and 40.8%, respectively), which was consistent with the results of droplet coverage and deposition. Among the other UASSs, pesticide utilization rates for JX25, DP, E-A2021, and T1000 were 35.4%, 35.9%, 34.9%, and 32.5%, respectively—5.4%, 6.0%, 8.3%, and 8.7% lower than the T20. The relatively low utilization efficiency of E-A2021 and T1000 may be attributed to E-A2021’s use of centrifugal nozzles, and T1000’s use of higher operational parameters. Overall, the pesticide utilization rate rankings of the UASSs were as follows: T20 > JX25 > DP > E-A2021 > T1000. When JX25, DP, E-A2021, and T1000 were sprayed with the adjuvants Nongjianfei, Silwet, Yimanchu, and Chengji, respectively, the pesticide utilization rates increased to 40.0%, 46.4%, 46.8%, and 39.6%, respectively, which were increased by 4.6%, 10.5%, 11.9%, and 7.1%, respectively. Notably, Yimanchu and Silwet showed the best performance, and the pesticide utilization rate was greatly improved after adding them, making E-A2021 spraying 5A·E and Yimanchu, and DP spraying 5AVM and Silwet the highest group.

The droplet ground loss rates for different treatments are illustrated in the scatter plot in Figure 9. The results indicated that droplet ground loss from UASS-based pesticide applications ranged from 34.5% to 46.8%, with no statistically significant differences (p = 0.2996). This overall trend was consistent with the pattern of pesticide utilization efficiency, suggesting a direct correlation between ground loss and effective application. This observation aligns with the findings of Qin et al. [40], though greater variability in ground loss rate was observed under wind disturbances and undulating ground. Without the addition of adjuvants, T20 spraying 5A·E and 5AVM resulted in slightly higher ground loss rates (41.5% and 40.9%, respectively) than when spraying 1.8AVM and 45B·E (40.1% and 38.0%, respectively). The ground loss rates for JX25, DP, E-A2021, and T1000 were 38.1%, 42.4%, 36.8%, and 43.5%, respectively. Compared to T20, JX25, DP, and T1000 showed increases of 0.1%, 1.5%, and 3.4%, whereas E-A2021 exhibited a 4.7% reduction. Among the five UASS models, the ranking of ground loss rates was: T1000 > DP > JX25 > T20 > E-A2021. T1000, which exhibited the lowest pesticide utilization rate and the highest ground loss, highlights the substantial influence of operational parameters on effective droplet deposition and loss. Excessively high flight parameters are unfavourable for enhancing deposition efficiency in mountain orchard operations. In contrast, the E-A2021 demonstrated both lower pesticide utilization rate and ground loss, suggesting that although its fine and dense droplets can penetrate the canopy and reduce ground loss, they are not conducive to improving pesticide utilization. When adjuvants were added, JX25, T1000, and DP—with the inclusion of Nongjianfei, Chengji, and Silwet, respectively—exhibited ground loss rates of 42.3%, 46.8%, and 44.6%, representing increases of 4.2%, 3.4%, and 1.5% compared to non-adjuvant treatments. Notably, E-A2021 combined with Yimanchu showed a reduced ground loss of 34.5%, a 2.3% decrease. Among all tested adjuvants, Yimanchu demonstrated the most promising performance in both enhancing pesticide utilization and reducing ground loss, followed by Silwet, Chengji, and Nongjianfei.

3.2. Control Effects of Different Spraying Methods

As shown in Figure 10, the control effect of five UASS models and the ground-based sprayer (ESG) against P. citri exhibited marked differences from 1 to 7 days after application (DAA). The control effects observed at 1 and 3 DAA were used to assess the rapid efficacy of different application methods. Among the four pesticides tested, 5AVM and 1.8AVM reached their peak control effect by 7 DAA, generally exceeding 92%, consistent with the rapid knockdown characteristics of AVM [10]. For both ESG and T20, 5AVM and 5A·E demonstrated the most rapid control effect among the four tested formulations, with very few live P. citri observed at multiple sampling sites (as shown in Figure 11b). Notably, ESG achieved over 90% control just 1 DAA, while T20 exceeded 90% control at 3 DAA. In contrast, 1.8AVM exhibited moderate speed of action, and 45B·E showed the slowest onset of efficacy. By 7 DAA, both T20 and ESG achieved over 90% control across all four pesticides, with 5AVM maintaining the highest efficacy. At this stage, citrus leaves had largely recovered, and live P. citri individuals were rarely detected (Figure 11c). Across the five UASSs, control effects remained lower than those of ESG at 1–7 DAA. Notably, at 1 DAA, DP, E-A2021, JX25, and T1000 exhibited significantly lower control effects than ESG (F = 22.5, p < 0.001), while T20 was the only UASS that approached ESG performance. Further comparison of different UASSs applying the same pesticide revealed T20 to deliver the best overall control effect. Specifically, compared to T20, the control effect of DP, E-A2021, JX25, and T1000 at 1 DAA was reduced by 1.1%, 4.8%, 5.0%, and 15.4%, respectively; at 3 DAA by 0.2%, 2.5%, 3.5%, and 10.9%; and at 7 DAA by 0.8%, 1.7%, 1.8%, and 5.4%, respectively. Interestingly, although E-A2021 presented the lowest droplet coverage and deposition, its rapid efficacy of spraying 5A·E exceeded that of JX25 and T1000. This suggests that finer, denser droplets produced by centrifugal nozzles may have advantages in targeting small-bodied pests like P. citri [26]. After adjuvant incorporation, control effects improved across all treatments. At 1 DAA, the control effects increased by 5.1%, 8.0%, 5.6%, and 8.3% with the use of Nongjianfei, Silwet, Chengji, and Yimanchu, respectively; while at 3 DAA, the improvements were 3.6%, 6.4%, 6.1%, and 9.6%, respectively. These findings highlight the superior performance of Yimanchu and Silwet in enhancing the early control effect, potentially by accelerating pesticide uptake and distribution.

Figure 10. Control effects at 1, 3, and 7 DAA.

Figure 10. Control effects at 1, 3, and 7 DAA.

Figure 11. Taking T20 spraying 5A·E as an example, the spraying effect and control effect before and after spraying. (a) P. citri occurrence on leaves before spraying; (b) distribution of leaf droplets and occurrence of P. citri at 1 DAA; (c) P. citri almost impossible to find at 7 DAA.

Figure 11. Taking T20 spraying 5A·E as an example, the spraying effect and control effect before and after spraying. (a) P. citri occurrence on leaves before spraying; (b) distribution of leaf droplets and occurrence of P. citri at 1 DAA; (c) P. citri almost impossible to find at 7 DAA.

As shown in

Table 4. Control effect of different treatment at 14 and 25 DAA.

Test	Spray Equipment	Testing Pesticides	Adjuvants	Control Effect 1 (%)
				14 DAA	25 DAA
1	JX25	45B·E	/	96.8 ± 3.5 abc 2	91.4 ± 5.8 abcd
2			Nongjianfei	98.7 ± 1.4 a	94.8 ± 4.8 ab
3	DP	5AVM	/	86.8 ± 5.3 f	80.9 ± 7.8 ef
4			Silwet	92.9 ± 4.1 bcd	86.3 ± 5.2 bcde
5	T1000	1.8AVM	/	87.3 ± 8.6 ef	79.8 ± 9.3 ef
6			Chengji	89.7 ± 5.3 def	84.2 ± 4.9 def
7	E-A2021	5A·E	/	96.8 ± 1.7 abc	91.6 ± 5.0 abcd
8			Yimanchu	99.2 ± 1.5 a	97.4 ± 3.7 a
9	T20	45B·E	/	95.3 ± 4.9 abc	89.7 ± 5.7 abcd
10		5AVM	/	88.3 ± 7.8 def	83.8 ± 9.2 def
11		1.8AVM	/	89.3 ± 5.1 def	85.8 ± 5.6 cdef
12		5A·E	/	96.3 ± 3.4 abc	93.6 ± 5.3 abc
13	ESG	45B·E	/	98.4 ± 2.2 ab	88.1 ± 11.6 bcde
14		5AVM	/	92.6 ± 9.4 cde	77.6 ± 17.5 f
15		1.8AVM	/	89.4 ± 9.0 def	86.3 ± 8.8 bcde
16		5A·E	/	97.6 ± 3.2 abc	86.8 ± 12.2 bcde

Note: 1 Data are shown as the means ± SD. 2 The control effect was performed by anti-sine transformation, and then variance analysis was performed. Different lowercase letters indicated significant differences at the p < 0.05 level (Duncan’s multiple range test).

, the control effect observed at 14 and 25 DAA exhibited patterns distinct from those observed before 7 DAA. The control effect at 25 DAA serves as an indicator for assessing the residual efficacy of different application treatments against P. citri. At 14 DAA, both 5A·E and 45B·E reached peak control levels (>95%), which is consistent with the mode of action of etoxazole and bifenazate—slow-acting but providing extended residual efficacy due to their interference with mite developmental processes [7,9]. In contrast, the control effect of 5AVM and 1.8AVM declined at 14 DAA, though both remained above 86%. When applying the same pesticide, UASS treatments generally exhibited lower control effects than the ESG. However, JX25 and E-A2021, applying 45B·E and 5A·E, respectively, achieved control effects of 96.8%, surpassing T20, while other UASSs still performed slightly below T20, with no statistically significant differences observed. At 25 DAA, the control effect declined across all treatments, with fewer than 10 live mites observed per sampled tree. The overall ranking of pesticide efficacy remained unchanged, with 5A·E and 45B·E outperforming 5AVM and 1.8AVM. Notably, 1.8AVM showed a pronounced drop in efficacy, reflecting the limited residual persistence of AVM under field conditions [10]. ESG showed slightly better residual performance with 45B·E than with 5A·E, while UASS treatments demonstrated the opposite pattern, indicating that the ME formulation may offer enhanced residual stability when applied via UASSs. Importantly, with the exception of T20 spraying 1.8AVM and E-A2021 spraying 1.8AVM, all other UASS treatments achieved control levels exceeding ESG at 25 DAA, suggesting that low-volume UASS spraying may confer advantages in long-term efficacy over high-volume ground applications [40]. However, Qin et al. also pointed out that under conditions of severe pest outbreaks, fluctuating wind environments, and tall or densely closed crop canopies, the long-term efficacy of UASS-based spraying may be inferior to that of conventional high-volume ground applications. Among UASSs, JX25 exhibited slightly superior residual performance with 45B·E compared to T20. Additionally, T20 achieved higher residual efficacy than T1000, DP, and E-A2021 by 6.0%, 2.9%, and 1.9%, respectively, when applying 1.8AVM, 5AVM, and 5A·E. The incorporation of adjuvants further enhanced residual efficacy across all treatments. At 25 DAA, the control effects’ improvements achieved by the addition of Nongjianfei, Silwet, Chengji, and Yimanchu were 3.4%, 5.4%, 4.4%, and 5.8%, respectively, with Yimanchu showing the most notable performance enhancement.

Among the four tested formulations, 5A·E and 5AVM demonstrated the most pronounced rapid efficacy, while 5A·E and 45B·E exhibited superior residual efficacy. The performance of 1.8AVM was relatively balanced in both short- and medium-term efficacy, albeit slightly inferior overall; thus, increasing its application rate may enhance its effectiveness. Across all UASS-based treatments, 5A·E emerged as the most effective formulation in terms of both rapid and residual control efficacy. With respect to UASS platforms, the rapid efficacy ranked in the order of T20 ≈ DP > E-A2021 ≈ JX25 > T1000, while the ranking for residual efficacy was JX25 ≈ T20 > E-A2021 ≈ DP > T1000. However, due to variations in pesticide formulation and operational parameters, this study does not attempt to directly compare the efficacy of identical formulations across UASS types. The enhancement effects of the four adjuvants on pest control followed the order of Yimanchu > Silwet > Chengji > Nongjianfei, which is consistent with the trends observed for droplet deposition and coverage.

In summary, although pest control in low-canopy, mountainous orchards remains challenging—particularly for P. citri—UASS can achieve effective management when operated under optimized parameters such as high spray volume, low flight speed, and low altitude, combined with appropriate pesticide and adjuvant selection. In recent years, several researchers have conducted empirical studies to validate and improve these application strategies [14,41]. For instance, in addressing more difficult-to-control diseases, Anken et al. [29] conducted trials in vineyards and found that when using the T50 for the control of grape downy mildew, UASS treatments achieved significantly higher efficacy on leaf surfaces compared to ground-based equipment. However, control efficacy on grape clusters was notably lower, indicating that while UASS is highly effective for foliar disease management, achieving uniform deposition in densely shaded canopy regions remains a technical challenge requiring further investigation.

4. Conclusions

This study evaluated the spray deposition performance and field-control effects of five different types of UASSs in complex mountainous citrus orchards for the management of P. citri. The results demonstrated that the T20 achieved the highest droplet coverage and deposition. Although the E-A2021, equipped with centrifugal nozzles, produced finer and denser droplets, its overall coverage and deposition efficiency were compromised due to increased drift potential. Nonetheless, it exhibited the highest inward and Inward Penetration Ratios (IPR and DPR). The JX25 and DP exhibited balanced performance across spray metrics. The T1000, operating under higher flight parameters for safety in mountainous terrain, showed relatively low deposition; however, the addition of adjuvants significantly improved its spray effectiveness. Regarding formulation types, pesticide formulation had a relatively minor impact on spray performance. ME outperformed EC, with SC performing the worst. All four adjuvants tested improved both droplet coverage and deposition to varying extents, with Yimanchu and Silwet showing the greatest synergistic effect, followed by Chengji and Nongjianfei. The overall pesticide utilization efficiency among the five UASSs ranged from 32.5% to 46.8%, with the T20 achieving the highest and the T1000 the lowest. Adjuvant addition improved utilization rates by 4.6–11.9%, although it also increased ground loss rates by 1.5–4.2%. Notably, Yimanchu improved utilization rate while reducing ground loss by 2.3%, indicating its superior drift control and retention performance.

In terms of control effects, UASS-based applications—while initially inferior to ground-based equipment (ESG)—surpassed ESG after the inclusion of adjuvants, particularly in long-term persistence. Among all treatments, the combination of E-A2021, the 5A·E formulation, and the Yimanchu adjuvant yielded the best short- and long-term control, maintaining an efficacy of 97.4% at 25 DAA. In short-term evaluations (1–7 DAA), the T20 demonstrated the most rapid control, followed by the DP. Although the E-A2021 showed relatively lower coverage and deposition, its fine droplet spectrum effectively targeted small-bodied pests such as P. citri, resulting in better short-term control than the T1000 and JX25. Over the medium-to-long term (14–25 DAA), the JX25 with 45B·E SC exhibited the most sustained control, followed by the T20. Formulation wise, both 5A·E and 45B·E showed significantly better persistence than AVM treatments, with adjuvants—particularly Yimanchu—substantially extending the duration of control.

Overall, this study confirms that UASS spraying, when optimized, can match or even outperform conventional ground-based equipment for P. citri management. However, optimal performance requires the integration of centrifugal atomization technology, appropriate adjuvant selection, and low-altitude, low-speed flight parameters to enhance canopy penetration and minimize drift and loss caused by terrain variation and environmental disturbances. Future research should focus on integrating multi-sensor feedback systems, improving droplet targeting strategies, and elucidating the synergistic mechanisms between pesticide formulations and adjuvants to maximize the efficacy and value of UASS-based plant protection.