Spatially Distributed Wind Measurements in the Atmospheric Boundary Layer with a Fleet of Quadrotors: Examination of the Homogeneity Assumption and the Applicability of Taylor’s Hypothesis

Spatially distributed wind measurements were conducted in the atmospheric boundary layer with a fleet of quadrotor UAVs (Unmanned Aerial Vehicle) at the Grenzschichtmessfeld (GM) Falkenberg during the FESSTVaL (Field Experiment on Sub-Mesoscale Spatio-Temporal Variability in Lindenberg) campaign in summer 2021. The big advantage of deploying the UAV fleet is the possibility to record the wind field at different spatial points simultaneously with a high temporal resolution. A particular flight pattern was performed to calibrate the UAV wind measures using reference data of the 99 m mast at GM Falkenberg. The results show that not only the mean wind but also the turbulence can be reasonably resolved up to 1 Hz. In a second pattern, ten UAVs were placed in a horizontal plane in both the streamwise and lateral directions of the flow. To align the pattern along the wind direction, easily adaptable flight plans were created and the last available wind measurements of the mast were used for the orientation. Varying spatial distances between neighbouring UAVs in this horizontal measurement pattern allow turbulence analyses as a function of distance. In this work, the spatially distributed measurements are used to examine the homogeneity assumption and Taylor's hypothesis in atmospheric boundary layer flows. The measurement strategy was repeated for a near neutral, a convective and a stable atmospheric boundary layer to identify possible dependencies of these assumptions on different atmospheric conditions. Considering the measurement accuracy of the UAVs, the results show no significant violation of the homogeneity assumption within the microscale with a maximum separation distance of 205 m. Calculated cross-correlations and coherences between pairs of UAVs aligned in streamwise direction reveal a decrease with distance, however. This indicates that Taylor's frozen turbulence hypothesis is not unrestrictedly valid for all scales. The strongest decrease with distance is observed for the stably stratified case, followed by the unstably one. The lowest decrease is found for the near neutral stratification. The measurement accuracies of the UAVs and a slight misalignment of the measurement pattern with the actual prevailing wind direction during the flight have to be considered, however, which are partly responsible for the decline. For the investigated separation distances from 5 m to 205 m, the achieved measurement duration of about 14 min is sufficient for the analyses of cross-correlations and coherences. This duration is determined by the battery runtime. However, the potential to resolve larger scales has already been demonstrated in the field by replacing the UAVs with charged ones during the flight