Long-range Airborne Water Vapour Lidar Observations: Comparisons with ECMWF Analyses and First 4D-VAR Assimilation Experiments

Kurzfassung

Airborne differential absorption lidar (DIAL) measurements of mid-latitude, subtropical and tropical water vapour were performed during several recent campaigns over North and South America, and Australia, as well as on the long-range transfer flights from and to Germany with the DLR “Falcon” research aircraft. These measurements include tropospheric as well as lower stratospheric observations. Complex advection and rapid vertical transport associated with meso- and synoptic scale dynamical processes reflect in inhomogeneous water vapor distributions along the flight paths, as observed by the lidar. The DIAL vertical H2O cross sections are compared to operational ECMWF analyses at T511/L60 and T799/L91 resolution, interpolated in space and time to the flight path. Qualitatively, the tropical (Hadley) circulation, stratospheric intrusions and humidity transitions between different air masses are analyzed by ECMWF with remarkable accuracy. Locally however, relative shifts and differing shapes of structures, particularly those exhibiting large q-gradients and being subject to rapid temporal development may cause considerable deviations. The relevant transport processes are adequately represented in the ECMWF analyses as long as vertical and horizontal scales exceed roughly 0.5-1 km in the vertical and roughly 50 km horizontally. Recently, impact studies at ECMWF started to investigate possible analysis and forecast improvement by assimilating these water vapour observations. First experiments indicated a reduction of the first-guess (12 h forecast) error compared to a control run, and the investigation of background and analysis departures confirmed the high quality of the data set. The basic DIAL processing steps and characteristics of the instrument errors will be presented. Further experiments to assimilate the full tropospheric and stratospheric lidar data set and sensitivity studies to optimize the use of the data in the ECMWF 4D-Var system are ongoing.