On the representation of frequency spectra in meteorology

Abstract

The common representation of frequency spectra in meteorology and climatology is discussed. It is pointed out that this representation is misleading since ‘spectral peaks’ and ‘spectral gaps’ are obtained even when the spectrum density is monotonously decreasing in the whole frequency range. A plea is made for using the spectrum distribution function, F(Ω) which gives an unambiguous picture of the distribution of variance with frequency.