Field comparison of spray drift from unmanned aerial spraying systems using air-induction nozzles versus rotary atomizers

Katrin Ahrens^*Andreas HerbstJens Wegener

• Julius Kuehn Institute Institute for Application Techniques in Plant Protection, Braunschweig, Germany

Unmanned aerial spraying systems (UASS, hereafter 'drones') are increasingly deployed for pesticide application, yet field evidence on spray drift—particularly for rotary atomizers (RAT)—remains limited. We conducted ISO 22866 drift trials over short grass using two drones: a DJI Agras T30 fitted with air-induction nozzles (AIN) and a DJI Agras T25 fitted with rotary atomizers. Droplet-size spectra were determined in accordance with ISO 25358 for AIN and via an adapted scanning procedure for RAT. Target application rate was ~75 L ha⁻¹ at a working width of 4 m, flight heights of 1.5–3.0 m, and flight speeds of 2.5– 4.5 m s⁻¹. Drift sediment was sampled 3–20 m downwind and expressed as a percentage of the application rate. Quantile regression (τ = 0.5, 0.9) on log-transformed sediment tested effects of distance, flight height, flight speed, atomizer type, and their interactions. Distance, flight height, atomizer type, and atomizer-specific interactions with flight speed and flight height were significant for the median response. RAT tended to produce higher drift under more extreme operating settings; however, at practical settings (2 m height, 2.5 m s⁻¹ speed) RAT drift was equal to or lower than AIN, with only small and distance-specific median differences. These findings indicate that atomizer choice and operating envelope jointly determine UASS drift. We discuss safe operating windows, limitations of RAT droplet sizing relative to ISO 25358, and provide open data and code to support meta-analysis and modeling.