Abstract
A currently used objective analysis scheme was tested by comparing a known function to an objective analysis of its values at the locations of radiosonde observing stations. A surface spherical harmonic of wavenumber 2 produced little spectral distortion, and 87% of the input amplitude squared appeared in wavenumber 2 in the objectively analyzed field. In contrast, only 13% of the input amplitude squared for a function of wavenumber 12 survived the analysis to appear in wavenumber 12. These results raise the possibility that spectral analyses of the kinetic energy in the atmospheric wind field may more properly be described by a less steep drop-off rate, for example, −2 rather than the −3 power law value measured from spectra derived from objectively analyzed date.
