Stephan Adam, Insititute of Physics and Meteorology, University of Hohenheim, Stuttgart, Germany
First Assimilation of Rotational Raman Lidar Temperature Data Into WRF
CoauthorsThomas Schwitalla, Eva Hammann, Andreas Behrendt, Volker Wulfmeyer
Abstract: For the first time, the impact of assimilating lower-tropospheric lidar temperature profiles was investigated. The profiles were measured with the Temperature Rotational Raman Lidar (TRRL) of the University of Hohenheim on a cloud free day with a well-developed convective boundary layer. The WRF model was used with 57 vertical levels covering most of Europe with 3 km grid increment. Three different experiments were carried out with a RUC with hourly 3DVAR assimilations. EXP_1 was performed with the assimilation of conventional data and the additional assimilation of TRRL profiles in a height region from 450 to 750 m up to 3000 m. In EXP_2, only conventional data were incorporated. The TRRL profiles were averaged over one hour, what lead to a statistical error of 0.01 K at 700 m increasing to 1.2 K at 3000 m. To consider representativity, an observation error of 0.7 K was chosen for the full height of the TRRL profiles. They were assimilated with the radiosonde operator at a height resolution of about 100 m.
The additional assimilation of the TRRL data corrected the temperature profiles as expected towards the lidar data. The temperature profiles in EXP_1 showed a root mean square error (RMSE) of only 0.6 K compared to the TRRL data, while the RMSE of EXP_2 was twice as large. Data of four radiosondes launched at the TRRL site were also available. Compared to these, the RMSE of the temperature profiles in EXP_1 and EXP_2 were 0.8 K and 0.7 K, respectively. The radiosonde profiles showed an overall RMSE of 1.1 K with respect to the TRRL profiles, which demonstrates the difficulties with radiosonde representativeness in the lower troposphere. The additional assimilation of TRRL data improved the boundary layer height and the temperature gradient in the interfacial layer in EXP_1 significantly. Thus, we can conclude that the TRRL data assimilation has great potential to close the critical gap of temperature observations in the lower troposphere.