For the coming years, a strongly growing use of unmanned aerial vehicles (UAV) of all classes is predicted, for example in agriculture, construction or disaster control. But, in order to quantify the actual traffic volume under different scenarios, appropriate models and empirical values are currently still lacking. These are of great use in research, as they make it possible to design regulations, technical support systems and airspaces in such a way that safe and efficient regular operation of UAVs can be ensured. Therefore, one of our SESAR projects DACUS is considering, among other things, how weather influences affect the capacity of urban airspace and operations. However, detailed micro weather forecasts today are not practical for a regular computation and a day-by-day usage in operation, since the calculation effort is too large and time-constrained. Instead, we will investigate in this project proposed here if it is possible to use pre-processed wind field data to train models that refine existing meso scale weather forecasts. Thus, we use a flow simulation to calculate wind speeds in an urban environment for typical wind directions and strengths identified in an observed area. Of special interest are the complex phenomena that exist in the so-called Urban Canopy Layer (UCL) and typical flight altitudes of UAV (500 ft. AGL). As a conclusion of the project, we envision an appropriate form of data preparation and presentation that ensures subsequent use in data-based learning procedures and the estimation of airspace capacity through risk zones.