Validation of the spray drift modeling software AGDISPpro applied to remotely piloted aerial application systems

Authors: CASTRO-TANZI, SEBASTION - Stone Environmental Consulting; WINCHEL, MICHAEL - Stone Environmental Consulting; TANG, ZHENXU - Bayer Corporation; TESKE, MILTON - Continuum Dynamics Inc; WHITEHOUSE, GLEN - Continuum Dynamics Inc; Fritz, Bradley; Martin, Daniel

Submitted to: Science of the Total Environment
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/2/2025
Publication Date: 2/8/2025
Citation: Castro-Tanzi, S., Winchel, M., Tang, Z., Teske, M., Whitehouse, G., Fritz, B.K., Martin, D.E. 2025. Validation of the spray drift modeling software AGDISPpro applied to remotely piloted aerial application systems. Science of the Total Environment. https://doi.org/10.1016/j.scitotenv.2025.178725.
DOI: https://doi.org/10.1016/j.scitotenv.2025.178725

Interpretive Summary: The growing use of spray drones and the need to better understand their off-target spray drift risk necessitates some method of accurately predicting the transport and fate of applied sprays. A field study was conducted using a commercial spray drone to develop a dataset for validation of a newly developed drift model. The work found that the model tended to under predict off target deposits due to the uncertainty associated with the application's effective spray swath width and the downwind displacement of the depositing material due to wind. The results are being used to refine the model's performance to provide users a tool to better understand and optimize application parameters, increase on-target deposits, and reduce environmental impact.

Technical Abstract: Given many operational advantages and the associated occupational health benefits, pesticide spray applications using unmanned aerial systems (UAS) are getting more attention in many countries. Regulatory frameworks evaluating environmental impacts of UAS pesticide applications are being actively developed by regulatory bodies worldwide. One important aspect of regulation development is to enhance the understanding of the off-target spray drift risk associated with pesticide spray applications using UAS. Currently, there are no validated mechanistic models available to simulate off-target deposition from UAS. To fill this research and regulatory gap, we evaluated a recently developed UAS drift model, AGDISPpro, adapted from AGDISP by incorporating several aerodynamical models of commercial UAS. The main goal of this study was to evaluate the performance of AGDISPpro for modeling spray deposition from UAS applications using two field studies with a focus on off-target spray drift. The field studies cover varied spray quality (nozzle size), meteorological conditions, and UAS operational factors. Study no. 1 was single-swath sprays using two different agricultural spray nozzles with Medium and Extremely Coarse spray droplet spectra. Study no. 2 was four swath sprays using Fine and Ultra Coarse spray nozzles. The ground deposition results predicted by AGDISPpro were compared to in-swath and off-target downwind field deposition measurements from the two field experiments and statistical analysis was performed to evaluate these comparisons. For study no. 1, the AGDISPpro simulation results compared better with the field observations for the Extremely Coarse DSD sprays than those for the Medium DSD sprays. The index of agreement between model predictions and field observations for off-target drift ranged from 0.47 to 0.92 (n=9) for Medium DSD sprays and from 0.61 to 0.94 (n=12) for the Extremely Coarse DSD sprays. Visual examination of the predicted deposition curves and field observations, along with prevailing negative mean bias error (d) values, indicate that AGDISPpro tended to predict slightly lower deposition values in the off-target section of the deposition profile compared to field observations. For study no. 2, AGDISPpro showed better performance in simulating field observations from the Fine sprays than those from Ultra Coarse sprays. The index of agreement between AGDISPpro predictions and field observations ranged from 0.86 to 0.93 (n=3) for the Fine sprays, and from 0.48 to 0.55 for the Ultra Coarse sprays. Similar to study no. 1, AGDISPpro also resulted in lower predictions of off-target spray deposition compared to field observations in most of the application events simulated. Uncertainty in UAS application swath width and swath displacement was identified as a component to the modeling requiring further investigation. A sensitivity analysis to determine the effect of swath width and swath displacement on uncertainty of model predictions indicates that these parameters greatly affect the magnitude of the resulting maximum peak deposition, and to a lesser extent, the width and position of the spray deposition plume. The analyses in this study demonstrate that AGDISPpro is a promising tool for modeling off-target spray drift deposition from UAS pesticide applications. Further research aimed at improving model prediction accuracy should include better understanding the determination of swath widths and swath displacements in UAS spray applications.