Electronic Thesis and Dissertation Repository


Doctor of Philosophy




Sica, Robert J.


An improved understanding of temperature variations in Earth’s middle atmosphere is important for the improvement of our understanding of climate and weather on the surface. The optimal estimation method (OEM) is an inversion modeling approach, which uses regularized nonlinear regression to retrieve, in this case, the temperature of Earth’s middle atmosphere using Rayleigh-scatter lidar measurements. The OEM regularization term is the a priori knowledge of the atmospheric temperature profile. In this thesis I use lidar temperatures in the altitude range 30–110km to construct a temperature climatology using over 500 nights of measurements obtained by the Purple Crow Lidar in London, Ontario. The OEM produces several diagnostic tools, such as averaging kernels and an uncertainty budget which includes both systematic and statistical uncertainties important for atmospheric applications. Using OEM allows for the quantitative calculation of the maximum valid altitude of the retrieval by determining at which altitude the a priori temperature profile influences the retrieval by

more than 10%. This new knowledge extends the temperature retrievals 5 to 10km higher in altitude than traditional methods. The OEM retrievals are validated by comparison of the PCL temperature climatology with other measurements. Excellent agreement is found between the PCL and sodium lidar climatologies in the upper mesosphere and lower thermosphere, where the temperature variability is highest. Thus validated, the OEM can now be applied to other similar lidar systems. Lidar retrievals of atmospheric temperature profiles using the OEM typically use a retrieval grid whose number of points is larger than the number of pieces of independent information obtainable from the measurements. Consequently, retrieved geophysical quantities contain some information from the a priori values, which can affect the temperatures at higher altitudes. I present a method for removing the a priori information from the retrieved profiles. The OEM provides averaging kernels, or

weighting functions, at each level. I applied the OEM to measurements obtained from two lidars during a coincident measurement campaign between the Deutscher Wetterdienst and National Aeronautics and Space Administration. The OEM averaging kernels are then used to improve lidar and satellite intercomparison.