Towards improved turbulence estimation with Doppler wind lidar VAD scans

Kurzfassung

Continuous profiling of turbulence parameters in the atmospheric boundary layer can be a major step towards improved numerical weather prediction models. Doppler wind lidars (DWL) have the potential to deliver the required data in operational observations. Velocity-azimuth display (VAD) type scans can be used to retrieve TKE dissipation rate through a fit of a turbulence model to the measured azimuth structure function. At the elevation angle of 35.3$^\circ$ it is also possible to derive TKE itself. We show in this study how modifications to existing methods allow to retrieve TKE and its dissipation rate even with a small number of scans, how a simple correction for advection improves the results at low altitudes and that VAD scans at different elevation angles with the same instrument provide comparable results of TKE dissipation rate after all filters and corrections. We show that especially for VAD scans with a narrow cone angle and at low measurement heights, the errors caused by advection without any correction are significant. Data of two experiments are utilized: First, measurements at the Meteorological Observatory Lindenberg (MOL-RAO) are used for validation of the DWL retrieval with sonic anemometers at two heights (50m and 90m) on a meteorological mast. Second, distributed measurements of three DWL in the Upper Silesian Coal Basin (USCB) during the CoMet campaign are analyzed and compared to in-situ measurements of the DLR Cessna Grand Caravan 208B aircraft (D-FDLR). In this experiment the VAD scans are performed successively at two different elevation angles: a narrow angle for wind profiles reaching up to 2000 m above ground and a VAD at 35.3° to retrieve TKE. The three lidar sites are separated by some tens of kilometers and represent areas of different terrain types. Looking at single campaign days, the analysis of turbulence shows that especially the lidar located in the fetch of a lake shows siginificantly lower boundary-layer heights and turbulence. Comparison of the lidar retrieved turbulence to in-situ measurements by D-FDLR reveal that the in-situ measurements above 800 m are higher than the lidar retrieval. In future, more comparisons of lidar turbulence retrievals to in-situ measurements (for example by aircraft or UAV) should be performed to better understand the uncertainties in these conditions