Development of Doppler radar wind data assimilation for the HIRLAM 3D-Var

M. Lindskog, H. Järvinen, D. B. Michelson

Abstract

A three-dimensional variational data assimilation (3D-Var) scheme has been developed for the HIgh Resolution Limited Area Model (HIRLAM). The HIRLAM 3D-Var minimizes a cost function measuring the distance of the analysis to the background field and to the observations. Each source of information is weighted according to its estimated error covariance. Radar wind observations can be input for HIRLAM 3D-Var either as radial wind super-observations or as vertical profiles of horizontal wind obtained with the Velocity Azimuth Display (VAD) technique. Background error standard deviation (sb) of wind and radial wind have been estimated by applying a randomization method, in which the background state is perturbed with Gaussian errors. The derived values of sb are used in the quality control of observations, and also in the assignment of radar wind observation error standard deviations (so). Parallel data assimilation and forecast experiments confirm reasonably well-tuned error statistics and indicate a slightly positive impact of radar wind data on the verification scores, regardless of whether the radar wind information is input as radial wind super-observations or as VAD wind profiles.

download report