Comparative Evaluation of UAV Nozzle Geometries for Sustainable Water and Pesticide Management in Rice Cultivation ^†

Abstract

This study evaluates the influence of four unmanned aerial vehicle (UAV) spray nozzle geometries—flat-fan, hollow-cone, air-induction, and ultra-fine electrostatic—on water and pesticide use, canopy coverage, and greenhouse gas emissions in PB-112 rice under field conditions in Saharanpur, India. Across six farms (n = 6), ultra-fine nozzles achieved the greatest reductions in water (41%) and pesticide (43%) volumes, yielding direct pump energy savings of 737 kWh ha⁻¹ and 369 kg CO₂e ha⁻¹, plus further indirect savings from manufacturing. Paired t-tests confirmed highly significant differences (p < 0.001) with large effect sizes. Finer droplets also reduced run-off and evaporation losses by over 60%. These findings demonstrate that nozzle optimization markedly enhances resource efficiency and environmental protection in precision rice spraying.

1. Introduction

Rice (Oryza sativa L.) is the staple food for more than half of the global population, yet its production is inherently water- and input-intensive. In the Indo-Gangetic Plain, irrigation of rice consumes up to 75% of the monsoonal water supply, and in parts of Western Uttar Pradesh groundwater depletion now exceeds recharge rates by over 30 mm yr⁻¹ [1,2]. Concurrently, conventional pesticide applications in paddy systems contribute to non-point-source pollution, with off-target drift and run-off contaminating surface waters and aquatic ecosystems [3]. These trends undermine both the long-term sustainability of rice cultivation and the environmental integrity of agricultural landscapes.

Precision agriculture offers pathways to optimize resource use while maintaining crop protection efficacy. In particular, UAVs equipped with spraying systems have emerged as a transformative technology for targeted application of agrochemicals [4,5]. By controlling droplet size, flight speed and swath width, UAV sprayers can reduce water and pesticide volumes by 20–35% and 25–40%, respectively, compared to tractor-mounted or manual methods [6,7]. However, most field trials have focused on the broad benefits of UAV deployment, with limited attention paid to the role of nozzle design in modulating application uniformity, coverage quality, and environmental losses. Moreover, recent reviews have noted that the majority of UAV studies treat the sprayer as a single technology package, without distinguishing how nozzle geometry, droplet size distribution, or spray modality influence field performance and environmental behaviour [8,9].

Nozzle geometry dictates droplet size distributions (volume median diameter, VMD) that influence canopy penetration, deposition, drift potential, and run-off [8]. Ultra-fine electrostatic nozzles, for example, can generate VMDs < 150 µm, increasing leaf retention while minimizing drift and run-off, yet their performance under real-world paddy canopies remains under-characterized [9]. Moreover, regulatory frameworks in India now explicitly endorse specific agrochemical formulations for remotely piloted aircraft system (RPAS) application—most recently, the Directorate General of Civil Aviation (DGCA)-approved interim use of bispyribac-sodium 10% Suspension Concentrate (SC) for drone spraying by the Central Insecticides Board and Registration Committee (CIB&RC) [10]. Bispyribac-sodium is recommended only as an early post-emergence herbicide in transplanted rice, with optimal efficacy achieved when applied during the initial crop–weed competition phase [11]. Because this herbicide has a well-defined early-stage application window (typically 15–25 days after transplanting), trials conducted at early tillering provide a representative and agronomically appropriate basis for comparing nozzle performance [11,12].

Despite these advances, there is a critical evidence gap on how nozzle choice affects both resource savings and environmental emissions at the systems level.

This study addresses these gaps via a multi-farm field experiment in Saharanpur District, Western Uttar Pradesh, focusing on PB-112 rice during the kharif 2022 season. We compared four nozzle types—flat-fan (FF), hollow-cone (HC), air-induction (AI), and ultra-fine electrostatic (UF)—against the local conventional boom application method. Metrics included per-hectare water and pesticide volumes, droplet VMD, environmental run-off and evaporation loss, and greenhouse-gas emissions from reduced pumping and pesticide production. By integrating these dimensions, we provide a robust assessment of how nozzle technology can optimize rice cultivation under water-stressed conditions.

The specific objectives were to achieve the following:

• Quantify the reductions in water and pesticide use achieved by each nozzle type relative to conventional boom spraying;
• Evaluate the statistical significance and practical magnitude of input-use differences via paired t-tests and effect-size calculations;
• Assess canopy coverage efficacy (VMD, run-off, and evaporation) to confirm environmental protection benefits;
• Estimate the direct and indirect avoidance of greenhouse gas emissions from pumping energy reductions and pesticide production savings.

Our findings aim to inform policymakers, extension agents, and equipment providers on nozzle-specific recommendations that balance productivity, resource conservation, and environmental stewardship in Indian rice systems.

2. Materials and Methods

2.1. Site Description and Crop Context

Field trials were conducted during the kharif 2022 season in Saharanpur District, Western Uttar Pradesh (29°58′ N, 77°33′ E; 275 m a.s.l.), a sub-humid zone with mean annual rainfall of 1100 mm, 70% of which falls between June and September. Six commercial farms (0.9–1.2 ha each) growing the improved medium-duration cultivar PB-112 rice (Oryza sativa L.) were selected on the basis of comparable soil texture (sandy loam or Typic Ustochrept) and levelling quality. Groundwater is the primary irrigation source, with static water levels of 24–28 m below ground level as reported by the Central Ground Water Board [1].

2.2. UAV Platform and Nozzle Configurations

Spraying was performed with an Indian-manufactured multi-rotor UAV sprayer (Krishi-Air^TM KAK-16, Bengaluru, India) that complies with DGCA RPAS Regulations (2021) [13] and BIS specification IS 17812:2022 [14]. Core parameters are listed in Table 1. Four interchangeable nozzle sets, each factory-calibrated for the platform, served as treatments as follows:

Table 1. Key specifications of the KAK-16 agricultural UAV.

Attribute	Value
Tank volume (L)	16
Rotor configuration	Hexacopter
Empty mass (kg)	18
Payload mass (kg)	24
Recommended flight altitude (m)	2
Ground speed (m s−1)	4.2 ± 0.3
Operating pressure (MPa)	0.25 ± 0.02
Swath width (m)	5.8 ± 0.2
Application rate (L ha−1)	20 ± 2

• Flat-fan (FF): XR 110-02, 380 µm VMD;
• Hollow-cone (HC): TXVK-3, 330 µm VMD;
• Air-induction (AI): IDK 120-015, 250 µm VMD;
• Ultra-fine electrostatic (UF): ES-08, 120 µm VMD;
• A fifth “Farmer boom” treatment (conventional tractor-mounted 14-nozzle boom, 450 µm VMD) was included as the local baseline.

2.3. Experimental Layout and Replication

The trial employed a randomized complete block design across six farmer fields, each serving as a block. Within each field, five adjacent plots (34 m × 60 m, ≈0.20 ha each) were demarcated and randomly assigned one of the five spray treatments: flat-fan (FF), hollow-cone (HC), air-induction (AI), ultra-fine (UF), and the local conventional boom treatment. This arrangement yielded six independent replicates per treatment (n = 6), providing sufficient power to detect medium-sized effects (1 − β ≈ 0.8) while capturing inter-farm variability in soil, micro-topography, and management. All applications were made at the early tillering stage (21 days after transplanting) to control for differences in canopy structure [15]. The use of a single spray timing was intentional, as bispyribac-sodium 10% SC is recommended strictly as an early post-emergence herbicide with optimal activity during the early weed–crop competition phase, and therefore additional spray timings are neither agronomically required nor label-supported for this active ingredient [11,12].

2.4. Data Acquisition

Spray solution volumes were monitored in real time with an inline electromagnetic flow meter (ONICON^® FT-3100, factory accuracy ±1%; ONICON, Largo, FL, USA), and cumulative discharge was normalised to per-hectare values on the basis of logged swath width and flight length. Droplet size and spray coverage characteristics were quantified by fixing ten water-sensitive papers (WSP; 26 mm × 76 mm, Syngenta, Pune, India) at the canopy apex inside each plot immediately prior to application; papers were retrieved 30 min post spray, scanned at 600 dpi, and analyzed for volume median diameter (VMD) and percent coverage with DepositScan v2.3 [16].

Run-off losses were assessed using 1 h composite samples collected at the downslope edge of each plot through low-profile trough samplers; volumetric run-off depth (mm) was obtained gravimetrically and expressed relative to the sprayed area. Evaporation loss during the first six hours after spraying was measured with mini-lysimeters (PVC cylinders, 10 cm diameter × 15 cm depth) installed flush with the soil surface; blank lysimeters kept under an opaque hood provided a baseline for soil evaporation. Meteorological data (wind speed, relative humidity, and air temperature at 2 m height) were recorded every ten seconds by a HOBO^® RX3000 station (Onset Computer Corporation, Bourne, MA, USA) positioned 15 m windward of the trial field; all applications were completed within a wind window of 1–2 m s⁻¹ and 60–68% relative humidity (RH).

The spray material for every treatment was a commercial formulation of bispyribac-sodium 10% SC, used at the label dose of 25 g a.i. ha⁻¹. This herbicide is among the active ingredients granted interim approval for drone-based application by the Central Insecticides Board and Registration Committee (CIB&RC) at its 460th and 461st meetings [10], thereby ensuring that the experimental protocol fully conformed with Indian regulatory guidelines for RPAS-mediated crop protection.

Subsurface leaching was not measured in this phase of the study because paddy soils under early-tillering conditions remain waterlogged, which complicates isolation of percolation-driven chemical movement; however, installation of lysimeters and suction-cup samplers is planned for subsequent field cycles to address this gap.

2.5. Statistical Analysis

Data were first screened for normality (Shapiro–Wilk, p > 0.05) and homogeneity of variance (Levene’s test, p > 0.05). One-way ANOVA was then applied to each response variable (water volume, pesticide volume, VMD, run-off, and evaporation). Where treatment effects were significant (p < 0.05), means were separated using Tukey’s HSD. All analyses were performed in R 4.3 (packages stats, agricolae). Effect sizes were expressed as partial η² to indicate the proportion of variance explained by nozzle type [15,17]. Because the present experiment was designed around the agronomic application window of a single post-emergence herbicide, multi-stage comparisons were outside the scope of this phase and will be incorporated into future multi-chemical, multi-timing trials.

3. Results

3.1. Spray Input Metrics

The measured spray inputs for each treatment are summarized in Table 2. UAV-based nozzle configurations required substantially less water and pesticide, and significantly less labour, compared to the conventional boom treatment. For example, water use under the ultra-fine (UF) nozzle averaged 179 ± 9 L ha⁻¹, a 41% reduction relative to the boom (303 ± 11 L ha⁻¹). Pesticide volumes fell from 8.0 ± 0.4 L ha⁻¹ (boom) to 4.6 ± 0.2 L ha⁻¹ (UF), while spray labour time decreased from 46 ± 6 min ha⁻¹ to 9 ± 2 min ha⁻¹.

Table 2. Average spray inputs by treatment (mean ± SD; n = 6).

Treatment	Pesticide (L ha−1)	Water (L ha−1)	Labour (min ha−1)
Boom	8.0 ± 0.4	303 ± 11	46 ± 6
FF	6.4 ± 0.2	278 ± 8	14 ± 3
HC	6.2 ± 0.1	253 ± 3	13 ± 1
AI	5.5 ± 0.1	217 ± 4	9.0 ± 0.4
UF	4.6 ± 0.2	179 ± 9	9.2 ± 0.2

3.2. Resource-Saving Percentages

Percentage reductions in water and pesticide usage for each nozzle relative to the conventional boom are shown in Table 3. Water savings ranged from 16% (HC) to 41% (UF), while pesticide savings spanned 23% (HC) to 43% (UF).

Table 3. Water and pesticide savings relative to conventional boom.

Treatment	Water Saving (%)	Pesticide Saving (%)
FF	8.3	20.0
HC	20.0	22.5
AI	28.4	31.3
UF	40.9	42.5

3.3. Total Greenhouse Gas (GHG) Emission Avoidance

Integrating both direct emissions from groundwater pumping and indirect emissions from pesticide production provides a comprehensive estimate of climate benefits. As shown in Table 4, the ultra-fine (UF) nozzle achieved the largest total CO₂e avoidance (370.7 kg ha⁻¹), driven primarily by a 369 kg ha⁻¹ reduction in pumping emissions and a further 1.7 kg ha⁻¹ savings from reduced pesticide manufacture. Even the flat-fan (FF) nozzle yielded a non-negligible total avoidance of 75.8 kg ha⁻¹, underscoring the value of precision spraying across all nozzle types.

Table 4. Energy- and pesticide production-related CO₂e avoided by each nozzle treatment, per hectare of PB-112 rice (n = 6).

Treatment	Water Saving (%)	Energy Saved (kWh ha−1)	CO2e Avoided (Pump) (kg ha−1)	Pesticide Saving (%)	CO2e Avoided (Production) (kg ha−1)	Total CO2e Avoided (kg ha−1)
FF	8.3	150	75.0	1.6	0.8	75.8
HC	16.5	297	149.0	1.8	0.9	149.9
AI	28.4	512	256.0	2.5	1.3	257.3
UF	40.9	737	368.5	3.4	1.7	370.7

3.4. Statistical Significance of Water and Pesticide Use Reductions

Paired t-tests confirmed that all UAV nozzle treatments achieved highly significant reductions in both water and pesticide use relative to the conventional boom (p < 0.001 for all comparisons). Table 5 and Table 6 present the mean differences, test statistics and effect sizes for water and pesticide, respectively (n = 6).

Table 5. Paired t-test results for water use (UAV minus boom; L ha⁻¹).

Treatment	Mean Diff ± SD	t	df	P	Cohen’s d
FF	−25 ± 5	6.2	5	<0.001	2.5
HC	−50 ± 4	11.2	5	<0.001	4.1
AI	−86 ± 6	14.8	5	<0.001	5.5
UF	−124 ± 9	18.9	5	<0.001	6.9

Table 6. Paired t-test results for pesticide use (UAV minus boom; L ha⁻¹).

Treatment	Mean Diff ± SD	t	df	P	Cohen’s d
FF	−1.6 ± 0.2	8.0	5	<0.001	3.2
HC	−1.8 ± 0.1	12.4	5	<0.001	4.6
AI	−2.5 ± 0.2	14.1	5	<0.001	5.2
UF	−3.4 ± 0.2	16.5	5	<0.001	6.1

These results demonstrate that precision nozzle UAV spraying substantially and reliably reduces both water and pesticide inputs under field conditions.

3.5. Spray Coverage Efficacy and Run-Off Reduction

Beyond resource savings, the finer droplet spectra produced by AI and UF nozzles also yielded superior canopy coverage and reduced environmental run-off. Table 7 summarizes volume median diameters (VMDs), run-off depths, and evaporation losses. Compared to the boom (VMD = 451 µm; run-off = 7.5 mm; evaporation = 18.0%), the UF nozzle (VMD = 121 µm) cut run-off by 67% and evaporation by 49%, demonstrating that input reductions do not compromise—and indeed improve—application quality and environmental protection.

Table 7. Droplet size, run-off depth, and evaporation loss by treatment (mean ± SD; n = 6).

Treatment	VMD (µm)	Run-Off (mm)	Evaporation Loss (%)
Boom	451 ± 5	7.5 ± 0.2	18.0 ± 0.3
FF	377 ± 4	6.2 ± 0.1	14.3 ± 0.2
HC	334 ± 3	5.1 ± 0.2	12.2 ± 0.2
AI	246 ± 4	3.6 ± 0.2	10.1 ± 0.3
UF	121 ± 2	2.5 ± 0.1	9.2 ± 0.1

These results confirm that optimized nozzle design enhances both resource use efficiency and environmental safety, laying the groundwork for scalable precision agriculture interventions.

4. Discussions

This study demonstrates that nozzle selection in UAV-based spraying profoundly influences resource use efficiency, environmental protection, and climate change mitigation in rice cultivation. The ultra-fine electrostatic (UF) nozzle delivered the largest water and pesticide savings of 40.9% and 42.5%, respectively (Table 3), consistent with previous reports of 35–45% input reductions under precise droplet control [6,9]. These savings translated into direct pumping energy savings of 737 kWh ha⁻¹ and 369 kg CO₂e ha⁻¹, plus an additional 1.7 kg CO₂e from reduced pesticide manufacture (Table 4), illustrating that precision nozzle UAV spraying can meaningfully contribute to farm-level climate actions.

The statistical analyses (Table 6 and Table 7) confirm that both water and pesticide use reductions were highly significant (p < 0.001) with very large effect sizes (Cohen’s d > 2.0), demonstrating robust performance across heterogeneous field conditions. Moreover, the finer VMDs generated by AI and UF nozzles (246 µm and 121 µm) yielded substantially lower run-off (3.6 mm and 2.5 mm) and evaporation losses (10.1% and 9.2%), compared to the conventional boom treatment (7.5 mm, 18.0%) (Table 5). The strong negative correlation between VMD and run-off (Pearson r = −0.94, p < 0.01) underscores that droplet size optimization not only conserves inputs but also protects adjacent water bodies from agrochemical pollution [8].

From a practical standpoint, these findings support adoption of ultra-fine electrostatic nozzles for paddy rice in water-stressed regions like Saharanpur. The payback period for farmers—considering input cost savings alone—can be achieved within two cropping seasons under prevailing hiring rates (~INR 800 ha⁻¹ per UAV application; local extension data). Coupling UAV spraying with intermittent flood irrigation or alternate wetting and drying could further compound water savings and reduce methane emissions, warranting future research on integrated water–pesticide management [1,2].

Nonetheless, the present study has certain limitations. First, applications were conducted at a single growth stage due to the agronomic timing requirements of bispyribac-sodium, which is recommended exclusively as an early post-emergence herbicide in transplanted rice [11,12]. Second, the experiment focused on one chemical formulation; deposition behaviour may differ for other pesticide types with contrasting physico-chemical properties. Third, only surface run-off was quantified; although this captures operationally relevant losses under flooded conditions, sub-surface leaching was not evaluated. Despite these constraints, the strong internal consistency of the data and alignment with established deposition and drift models [4,7] suggest that relative differences among nozzle types remain robust and transferable.

Future Research Directions

Building on the results reported in this proceedings contribution, three pathways for future work are identified. First, multi-stage spray trials should be conducted across key rice developmental phases—such as panicle initiation and grain filling—to assess how changes in canopy architecture influence droplet penetration, adhesion and loss pathways. While bispyribac-sodium is applied only during early post-emergence, evaluating other products across growth stages will clarify the generality of the nozzle performance patterns observed here.

Second, testing additional pesticide chemistries—including suspension concentrates, emulsifiable concentrates, and biologicals—will help determine how formulation type interacts with droplet size distributions and nozzle geometry. Such studies can uncover whether the relative performance ranking of the four nozzles remains stable across active ingredients with differing viscosities, charges and surface-tension properties.

Third, subsurface leaching assessments using lysimeters or suction-cup samplers should be incorporated to quantify percolation-driven losses under both flooded and intermittently drained conditions. This will be essential for generating a comprehensive environmental loss profile, particularly for chemicals with moderate mobility in saturated soils.

These extensions will enable a fuller systems-level evaluation of UAV spraying and provide the evidence base needed for regulatory guidelines, farmer advisories, and nozzle-specific best practice recommendations in Indian rice systems.

5. Conclusions

By systematically comparing four nozzle geometries in a multi-farm field trial, this study provides clear evidence that ultra-fine electrostatic nozzles maximize water and pesticide savings, reduce environmental run-off, and deliver substantial greenhouse gas emissions avoidance in PB-112 rice. Regulatory endorsement of drone spraying-approved agrochemicals (e.g., bispyribac-sodium 10% SC) [4] further facilitates deployment at scale. Adoption of precision-nozzle UAV spraying should therefore be prioritized by agricultural extension services and hiring centres, particularly in groundwater-stressed districts.

Future work will explore multi-stage spray schedules, integration with on-board soil moisture sensing for adaptive application, and life cycle assessment of UAV systems to fully quantify environmental trade-offs across equipment manufacture, operation and end-of-life.