Abstract (eng)
Data assimilation allows for the direct integration of meteorological observations into atmospheric models, significantly improving numerical weather prediction. This technique, along with the continued incorporation of new types of observational data, has led to significant advances. However, the assimilation of data and the generation of accurate forecasts in regions with complex topography, such as mountainous areas, remain particularly challenging. This thesis explores an innovative approach: the direct assimilation of Doppler LIDAR radial velocity measurements into a high-resolution weather model to improve the state estimation of the atmosphere, with a focus on valley circulation. Such a state estimation - viewed even as a local campaign data reanalysis can combine and supplement the information from campaign observations with physically consistent model simulations. The study uses observations from the CROSSINN campaign in the Inn Valley to refine the representation of meso and micro scale alpine atmospheric processes. These thermally driven winds are difficult to accurately represent and forecast in current numerical weather prediction models. This work examines whether a direct assimilation of radial velocity can provide a more accurate state estimation of the atmosphere compared with the conversion into vertical profiles of horizontal wind using the Velocity Azimuth Display (VAD) technique. It also evaluates the effectiveness of combining LIDAR data with operationally used sources such as ground stations, radiosondes, and air reports. The results show a complex picture. Depending on the validation method used, the direct assimilation of Doppler LIDAR data can lead to a reduction in analysis and forecast errors, thus improving the predictive accuracy of the model. In particular, the combination of multiple data sources produces the most robust results, highlighting the value of integrating diverse observational data to capture the full complexity of atmospheric conditions. This approach can not only improve the understanding and representation of thermally driven valley flow, but also provides valuable context for observations from the CROSSINN campaign and future observational efforts, potentially leading to more accurate weather forecasts in complex terrain regions.