Using the SWUF-3D drone fleet to determine turbulence characteristics in a small Alpine valley during TEAMx

Three-dimensional turbulent processes substantially contribute to the exchange of energy and momentum in the mountain boundary layer (MoBL). These processes are usually not properly represented in current numerical weather prediction models, since turbulence parameterizations are not adapted to complex terrain. To better understand the three-dimensional structure of turbulence in the MoBL and to inform future parameterization development, in-situ observational data from MoBLs is necessary. The optimal dataset would consist of 3D wind and standard atmospheric state variables measured at distributed locations simultaneously. The novel method of using a fleet of drones (i.e., Uncrewed Aerial Systems, UAS) can provide an approximation to this goal.

The drones of the SWUF-3D fleet allow for three-dimensional wind measurements and are equipped with fast temperature sensors and a capacitive humidity sensor. They can be programmed to hover at certain positions for a duration of approximately 20 minutes. Three measurement campaigns were conducted at Nafingalm, an Alpine pasture near Innsbruck, in the framework of the TEAMx research programme. The most recent campaign, carried out in summer 2025, encompassed a maximum of 30 UAS hovering between 50 and 240 m above ground level (AGL). Depending on the chosen pattern in which the drones are organized, different analyses can be performed. Turbulent fluxes and turbulence kinetic energy (TKE) can be compared between different drones for any pattern. Arranging the drones in a box-like pattern allows for computing gradients between UAS in all three spatial directions. All of the turbulence quantities to be analyzed require at least one of the three wind components. Deriving absolute vertical velocity accurately is particularly challenging. An essential parameter for vertical wind calculation is the rotor thrust of the multicopter. Thrust is directly proportional to air density.

Analysis showed that approximating the temperature by the international standard atmosphere to derive the density at the measurement site nearly 2000 m above mean sea level can lead to substantial errors in absolute vertical wind. In highly three-dimensional flow such as in an Alpine valley, an accurate estimate for the absolute vertical velocity is important to understand orographic flow deflection. Therefore, if UAS are used to obtain the 3D wind vector, an integrated sensor suite for temperature, humidity and pressure is essential and thermodynamic measurements need to be thoroughly quality-checked for all variables. Regarding TKE, preliminary results show that the drones reasonably capture the diurnal evolution when compared to a ground-based sonic anemometer located 5 m AGL. TKE and turbulent momentum fluxes at the different drone positions indicate spatial heterogeneity for given points in time. This contribution will focus on the technique of measuring turbulence with a drone fleet as well as the analysis of different turbulence quantities and their distribution across the valley.