Application of Unmanned Aerial Vehicles for Atmospheric Sampling: A Numerical Experiment By Large-Eddy Simulation

Volatile organic compounds (VOCs) play critical roles in ecosystem functioning, climate change, and atmospheric chemistry. Sampling of VOCs over forest at intermediate scales of a few hundred meters is challenging due to the absence of appropriate platforms. Unmanned aerial vehicles (UAVs) represent an emerging advance in atmospheric chemistry in this regard because of excellent maneuverability and sampling capability at these scales. In this study, numerical experiments based on large-eddy simulation were utilized to assess the role of turbulent transport of VOCs in the forest atmospheric boundary layer to provide guidance on atmospheric UAV VOC sampling. The numerical experiment mimics surface and boundary layer conditions similar to the ones observed and enable us to perform sensitivity conditions regarding the sampling conditions. Characteristic vertical profiles were explored through the simulations for chemical species of short (10² s), medium (10⁵ s) and long (10⁸ s) atmospheric lifetimes. Results demonstrate that near-canopy sampling is favorable for both short- and medium-lived because their concentrations drop dramatically with altitude. Lower concentrations of VOCs require longer sampling times for representative sampling, so as to average out the stochastic effects of atmospheric turbulence with time scales varying from a few seconds to several minutes. In general, a duration of greater than 5 min is recommended for VOC samplings to have < 20% uncertainty caused by atmospheric turbulence. A maximum sampling height of 200 m is proposed for UAV sampling over heterogeneous land surface in order to observe any horizontal emission variability. Based on the results, a model framework is developed to estimate the contribution of upwind emissions on the UAV sampling. Results from this study are focused on providing an understanding of VOC transport in the atmosphere in the context of effective UAV sampling at intermediate scales over forest.