Determining tidal-period perturbations from less than full-diurnal cycle time series

Lidar and radar observation can now be performed continuously for several complete diurnal cycles. Though diurnal and semidiurnal amplitudes are typically larger than the terdiurnal and quatradiurnal counterparts, depending on season and latitude, such is, however, not always the case. For the long-period continuous data sets decomposition into tidal-period oscillations with diurnal, semidiurnal, terdiurnal and quatradiurnal components is unique. Often, the data sets shorter than a full-diurnal cycle coverage are nonetheless of considerable interest, because of the need to assess the contribution of tidal-period perturbation to atmospheric instability or the tidal variability due to gravity wave interaction which may be very strong. Unfortunately, data shorter than one complete diurnal cycle admits diverse decomposition, depending how many terms are included in the analysis. In this paper, we use synthetic data with reasonable amplitudes to assess the variability or uncertainty of deduced tidal components due to different relative phases of the actual tides, due to different assumed components in the analysis, as well as due to the length of data sets and the presence of gaps. We then discuss these findings by investigating a 9-day continuous data set and two other sets each with 16 hours of data, all acquired in September by our two-beam lidar, capable of measuring temperature, zonal and meridional wind in the mesopause region.