Retrieval of Micro-scale Flow Structures from High Resolution Doppler Lidar Data Using an Adjoint Model

Radial velocity data measured by high-resolution Doppler lidar (HRDL) are assimilated with a four-dimensional variational data assimilation (4DVAR) method to retrieve microscale flow structures in an atmospheric boundary layer. The control variables of 4DVAR consist of initial three-dimensional velocity and temperature fields, as well as profiles of eddy viscosity and thermal diffusivity. The effects of lateral boundary conditions and buffer zones on the retrieval quality are evaluated using identical twin experiments with synthetic observational data prior to HRDL data retrieval. It is found that use of inflow/outflow boundary conditions in conjunction with buffer zones yields good results. HRDL field observations are then assimilated with 4DVAR, together with the above treatment of boundaries to recover microscale flow structures in a convective boundary layer. The uncertainty of the retrieved data is assessed by conducting a grid sensitivity test. A large-scale flow structure resembling a dry microburst is observed in the retrieved wind field. Characteristics of this structure are discussed and compared with those of a typical microburst.