From about 100 nights of ground-based measurements of fluctuations in airglow brightness and rotational temperature, taken from two sites at 32øS and 37øN, a considerable number of new determinations of complex values of Krassovsky's •1 are derived, and compared with observations by othe• investigators and model predictions.Our findings support the model of HineS, Tarasick, and Shepherd (HTS), which leads to practical consequences regarding the usefulness of airglow observations for deriving vertical propagation of atmospheric waves in the mesopause region. It is shown, at least in part, that the vertical wave propagation can 'be inferred from zenith observations of one airglow emission, alone, and that consistent information can be obtained simultaneously for the two airglow layers. The analysis presented comprises the range of observed periods between 3 and 24 hours. The salient feature is that •1 values are essentially limited to the fourth quadrant, for both the 02 and OH emissions, which means, according to the HTS model, that the majority of the waves observed propagate upward, with vertical wavelengths between 20 and 60 km, and only a few are possibly evanescent. This would not contradict the interpretation that most if not all of the wave signatures may be due to the semidiurnal tide, or tidal transients.
IntroductionThere is now general agreement that temporal variations of the airglow intensities and rotational temperatures are due to gravity waves passing through the rnesopause region. However, exactly how to extract dynamical information from airglow observations is far from clear. This is because a detailed quantitative description of how the airglow layers react to the passage of waves has not reached full consistency with observations. Existing models still disagree with each other in many respects, and the relatively few experimental results have not been sufficiently specific and consistent to show clearly if and how models should be improved.A key parameter is Krassovsky's rl, defined as the ratio of the relative amplitudes of the oscillations in airglow intensity and rotational temperature [Krassovsky, 1972]. This is nowadays taken to be complex. That is, it includes the phase shift between the intensity and temperature oscillations.According to the model of Hines, Tarasick, and Shepherd (HTS model) [Hines and Tarasick, 1987; Tarasick and Hines, 1990; Tarasick and Shepherd, 1992a, b], the imaginary part of rl gives information about the vertical propagation of gravity waves. Until now, however, this prediction has not been experimentally tested. The models of Walterscheid et al. [1987, 1994, and references therein] do not contain such a feature, at least explicitly. In two other models, by Zhang et al. [1993b], for the 0 2 emission, and by Makhlouf et al. [1995], for OH, it is also not clearly visible whether this feature is present. Several experimental determinations of rl have been more recently reported in the literature for the OH emission [Sivjee et al., 1987; Hecht et al., 1987; Swenson ...