Abstract: In order to effectively identify and extract Doppler weather radar wind field information,the dealiasing method of Doppler radar radial velocity data proposed by Gao and Droegemeier based on the variational technique was improved.In the original method,the information of background wind field,azimuth and radial velocity and distance gradient were simultaneously taken as constraint conditions to calibrate the Nyquist number.However,in the iterative process,a large number of numerical analysis and partial differential equation were used,resulting in the radial velocity excessively smooth and distortion.To solve this problem,based on obtaining the radial velocity analysis field,the improved algorithm combined with the original radial-velocity observation field,automatically identified the area with velocity ambiguity and calculated the Nyquist number to be corrected through the image change detection method,and then corrected the aliasing area of the observation field.The feasibility of the improved algorithm was verified by the radar volume scans data of severe convection weather and a typhoon event.Compared with the original and the operational WSR-88D algorithms,it shows that the improved algorithm can effectively solve the shortcomings of the original algorithm,reproduce the real structure and distribution characteristics of the wind field,improve the quality of velocity dealiasing,and thus obtain a more reasonable radial-velocity observation field.Moreover,the dealiasing effect of this algorithm is better than that of the WSR-88D algorithm,which is helpful for scientific research and operational application.