Combining ground-based and space-based data to study tides in the mesosphere and lower thermosphere

In this thesis I investigate the accuracies to which tides can be inferred in the mesosphere and lower thermosphere using satellite-based data. Satellite sampling scenarios affect the recovery of diurnal and semidiurnal tidal winds because of slow local time precession rates. Ground radars have good temporal coverage, but are limited to a finite number of locations. Combining ground-based data with satellite data enhances the local time coverage, reducing the number of days necessary to accurately retrieve tidal amplitudes and phases. The degree to which ground radar measurements, when combined with satellite data, improve the determinations of longitudinal tidal structures represents a major focus of this work. To this end, new methodologies are developed using the Kyushu University GCM to generate proxy satellite and ground data so that issues such as sampling, aliasing, intercalibration and ground data weighting can be examined. Conclusions drawn regarding these issues are applied to analyses of actual data from the UARS satellite instruments and ground radars.; The northern midlatitudes are examined using output from the GCM sampled according to the UARS and TIMED satellite instruments. The two satellites are characterized by different orbits which result in different precession periods. Experiments are designed to systematically explore weighting and aliasing over the course of each precession period while a true, or fully-sampled GCM, result is known. Ground station placement is important for overcoming aliasing errors, and applying a ground weighting factor of 0.5 to 2.0 improves the overall accuracy of the inferred tidal structures. A longitude variation in the semidiurnal tide near 55°N is discovered by actual measurement and the above methodology.; The tropical and equatorial regions are examined using the TIMED-sampled Kyushu GCM. TIDI uses a "dual-viewing" sampling, measuring two local times per latitude crossing. This configuration needs approximately 30 days of satellite data to infer the diurnal tides, which is further reduced to five days by adding ground data. SABER requires a full 90 days to correctly infer the diurnal tides.; Finally, the MLT tides over Antarctica are inferred. A new two-dimensional basis function is created, incorporating both latitude and longitude. Climatological UARS and radar data, and their GCM-sampled counterparts, are used to examine the two-dimensional structure of semidiurnal tides from 60°S to 90°S, and to produce the first semidiurnal tidal climatology over the Antarctic continent.