Structures in the Horizontal Wind Field over Urban Mountainous Terrain detected with Coplanar Doppler Lidar Scans

The flow in the atmospheric boundary layer over urban mountainous terrain is highly variable in space and time, as multi-scale processes such as convection and thermally and dynamically driven flows superimpose and interact. To capture this variability, area-wide observations with appropriate high temporal and spatial resolution are necessary. Traditional in situ measurements on towers are problematic in such a complex environment, as they e.g. are often not representative, limited by the inaccessibility of the topography and not possible at all above the buildings. Remote sensing with Doppler lidars is a promising approach to capture the large spatio-temporal variability of the flow in the atmospheric boundary layer above the topography. By performing synchronized horizontal scans with three Doppler lidars installed on slopes overlooking the city of Stuttgart, which is located in moderate mountainous terrain in south-western Germany, we retrieved the horizontal wind field in a horizontal plane of several square kilometers with a temporal resolution of 1 min and a spatial resolution of 100 m. The terrain around Stuttgart is characterized by a basin-shaped valley (Stuttgart basin) which opens into the larger Neckar Valley.

Using horizontal wind field data on 22 days in July and August 2018, we investigated on the one hand, the mesoscale structure of the horizontal flow in the valleys in dependence of time of the day, atmospheric stratification and the wind above ridge height. On the other hand, we analyzed how fast cells in the convective boundary layer propagate downstream, i.e. we estimated the convection velocity. During stable conditions, downvalley wind dominated in the valleys and the flow below ridge height was decoupled from the flow aloft. During thermally unstable conditions, the flow in the valleys is mainly coupled to the flow above ridge height with a preference for upvalley wind being visible. Cells moving downstream were detected in the convective boundary layer and we applied a method based on the two-dimensional auto-correlation to estimate the convection velocity. The convection velocity was found to be higher by 0.6 m s^-1 (24 %) than the mean horizontal wind speed at the height of the horizontal plane.