Time series of wind velocity measured with a Doppler lidar at different altitudes have been analyzed using the local coherence function of the wavelet transforms of signals (Izvestiya RAN, Atmospheric and Oceanic Physics 37, No. 5, 584-591 (2001)). Feasibility of isolating groups of wind data with the coordinated behavior is demonstrated using a particular measurement session. It is shown that the coherence function allows one to isolate 2-D time-frequency domains with the high correlation of the wind velocity data upon the preliminary compensation for the local phase shifts of the same scale. The approach considered can be used for the detection and study of the dynamics of coherent formations in the atmosphere.