Zonal mean and tidal dynamics from space: an empirical examination of aliasing and sampling issues

Interpretations of space-based measurements of atmospheric parameters in the mesosphere and thermosphere are complicated by large local-time variations at these altitudes. For this reason, satellite orbits are often preferred which precess through all local times one or more times per season. However, the local-time structure of the atmosphere is inherently non-stationary, which can lead to sampling and aliasing diculties when attempting to deconvolve the measurements into zonal mean and tidal components. In the present study, hourly radar measurements of mesopause-region winds are used to form a mock data base which can be used to gain insight into implications of the aforementioned problems; the use of actual measurements introduces a realistic element of geophysical temporal variability. Assuming zonal symmetry (i.e., migrating tides superimposed on a zonal mean circulation), the radar measurements are sampled from the satellite perspective for orbital inclinations of 57 and 70 , and compared to the ground or true perspective. These comparisons provide realistic estimates of the errors to be expected when attempting to derive mean and tidal components from space-based measurements. For both diurnal and semidiurnal components, and the quoted satellite inclinations, acceptable errors …3±4m/s rms† are obtained for data covering 24 h local time (i.e., ascending plus descending nodes); the corresponding errors for singlenode data (12 h local-time coverage) are of order 8±11m/s, and therefore may not represent reliable estimates of the actual tidal components. There exist certain caveats in connection with the latter conclusion which are discussed. 1 Introduction The feasibility of determining zonal mean and atmospheric tidal components of dynamical ®elds from a local-time precessing space platform is the subject of ongoing discussion and speculation. Several recent papers have appeared which derive tidal oscillations in the neutral wind in the mesosphere/lower thermosphere region (c. 80±120 km; Burrage et al., 1995; Mclandress et al., 1996) using observations from the Upper-Atmosphere Research Satellite (UARS) which has an orbital inclination of 57 and local-time precession period near 36 days. Similar studies will be performed using data from the TIMED (thermosphere-ionosphere-mesosphere energetics and dynamics) mission, which will have an inclination near 70 (http://sd-www.jhuapl.edu/ TIMED). The diculties involved fall into the categories of sampling and aliasing. The sampling issue centers around the inherent geophysical variability of the system within the 24-h local-time precession period of the satellite, which is about 60 days for an inclination i of 70 …ascending‡ descending node†. In other words, the atmospheric structure ordered in local time does not remain stationary while it is being sampled by the satellite instruments. The aliasing issue is somewhat di€erent. Let us suppose that the tidal components remain stationary, but the zonal mean varies with time. If the atmosphere is sampled from the satellite perspective under this scenario, as slow local-time precession takes place, then the changing zonal mean is perceived as a local-time change, leading to spurious tidal harmonics. In this case we say the mean zonal ̄ow aliases into the tidal components. Evolving tidal structures can alias into the mean, but as we will see, this is a much smaller e€ect, since the tides tend to cancel out when determining the mean. The present work takes an empirical approach to the problem, using actual mesopause-region hourly wind observations from the Christmas Island, Hawaii, and Correspondence to: J. M. Forbes Ann. Geophysicae 15, 1158±1164 (1997) Ó EGS ± Springer-Verlag 1997