風速変動のウェイブレット解析

Abstract

Wavelet analysis is a method of time series analysis based on an expansion in terms of tem-porally localized functions. The basis functions are called wavelets, and are constructed bytranslation and dilation of a single localized function called analyzing wavelet. The waveletshave two parameters; one specifies the time scale of variation and the other the time of events.The wavelet spectrum is directly related to the Fourier spectrum and enables us to analyze arelation between the characteristics of Fourier spectrum and temporally local events in the timeseries. Here the wavelet analysis is applied to atmospheric wind data and the power low of theFourier spectrum is discussed in relation to mean wind speed and environmental terrain.