APPLICATION OF WAVELET ANALYSIS TO WIND DISTURBANCES OBSERVED WITH MST RADAR TECHNIQUES

A computer program following the orthonormal wavelet analysis algorith m developed by Yamada and Ohkitani (1991) is applied to an analysis of local and transient behaviours of internal gravity waves from a finit e-length, discrete data record of the vertical profile of wind velocit y provided by the MST radar technique. The functionality of this progr am has been confirmed by simultaneously simulating well-known spectral and monochromatic features (such as the -3 power law of a vertical wa ve-number spectrum and the upward increase of predominant vertical wav elength) by a summation of several wavelet components corresponding to localized gravity waves. This simulation is used also to study the re liability of the wavelet analysis program. It is shown that, by adding 64 null data at each end of the profiles, all the wavelet coefficient s can be obtained by this program, apart from those at each end. We ha ve applied this program to about 1000 vertical profiles of zonal and m eridional winds in the troposphere and the lower stratosphere (2-20 km altitude) which are obtained from 30 min averages of three weeks of c ontinuous observation data with an MST radar (the MU radar in Japan) d uring June-July 1991. We find from the wavelet analysis that quasi-mon ochromatic waves with vertical wavelength approximate to 2 km are domi nant above the tropopause, and that activities of the shorter or longe r waves have different vertical distributions. Striking temporal varia tions of vertical distributions of wave activity are also clearly foun d by the wavelet analysis.