Wind flows inside complex urban environments are determined by the mutual effect of large-scale (i.e., wind above the urban boundary layer, UBL) and local-scale forcing (i.e. constraints into the urban canopy layer, UCL). Usually, when the wind field above the UBL is known (e.g. by on-site measurements, OsM), high-resolution microscale models (e.g. CFD) are used to predict the wind inside the UCL. However, standard procedures to map the wind field from an undisturbed position to the UCL are not available yet and several technical aspects need to be investigated. A downscaling method, so-called “static downscaling”, of the wind from mesoscale to microscale is innovatively adopted here to evaluate the performance of two CFD microscale models when predicting the flow in a UCL. The methodology is based on OsM transferred into the UCL by means of so-called “transfer coefficients” calculated by 3D steady RANS simulations for two different spatial extents of the explicitly modeled urban texture (Case A and B). It is discussed in detail the way the transfer coefficients work, how they can be used to understand the correlations between wind above and within the UCL as well as to identify the major limitations of the RANS approach. Results are discussed in qualitative terms and quantified using standard metrics. It is also shown that too high flow rates can occur at the entrance of the waterway of the area of interest and in the outer part of the explicitly modeled urban area, which can be mitigated by “buffer zones”. Display omitted •On-site measurements were performed in Livorno city with LiDAR and anemometers.•RANS simulations on more and less spatially extended cases of Livorno city.•Transferred CFD data were compared to measured data inside the UCL at target points.•The extension of the environment surrounding the target area plays a key role.•Deviations are discussed and effects on wind flow quantified by statistical metrics.