Spectra of Atmospheric Variables as Deduced from Ground-Based Radiometry

Ground-based radiometric observations have proven to be effective means for remotely determining both the static and the dynamic thermal vertical structures of the lower troposphere. Since the meteorological parameters are random fields, the atmospheric radiance measured by a ground-based radiometer fluctuates randomly in time, and under suitable conditions, these fluctuations result essentially from atmospheric temperature fluctuations. A relationship between the spectral density of the output of the radiometer and the spectrum of the atmospheric temperature is obtained, and in particular the special case of frozen turbulence is investigated. In the experiment which is reported, the downgoing radiance has been measured in several bands of the infrared in which the atmosphere exhibits different absorptions. The low-frequency spectral density of the fluctuating radiance has been computed both by a suitably windowed fast Fourier transform and by the maximum-entropy methods. The latter technique is shown to yield either high spectral resolution or enhanced smoothing, according to the order of the prediction filter which controls the spectral-estimation procedure. Data are presented on two classes of spectra corresponding to different stability conditions of the atmospheric boundary layer.