Tomographic reconstruction of the time-varying ionosphere

Computerized ionospheric tomography (CIT) is an inexpensive and effective way of obtaining information about the ionosphere. The central problem of ionospheric tomography is that the total electron content (TEC) data are not sufficient to determine a unique solution. In current algorithms for CIT reconstruction, some of the information in the TEC data is neglected and replaced with a priori information. This thesis will present two new algorithms for static CIT. In addition, current algorithms do not take into account time variations in the ionosphere during data collection. This thesis will present analysis of the effect of time variations in the ionosphere on static reconstruction, the first algorithm for time varying CIT, and analysis of the configuration of a time varying CIT system.; The two static algorithms to be presented are the residual correction method (RCM) and smoothness and conjugate gradients (SCG). Both algorithms are iterative. In RCM the solution is confined to a solution space consisting of reasonable solutions based on a priori information. SCG uses smoothness constraints to establish a unique solution within the set of solutions that satisfy the TEC data. Although RCM offers superior resolution when the ionosphere is consistent with the model, SCG provides an alternate way of performing reconstruction when the correctness of the model is in doubt.; The effects of time variations in the ionosphere on static reconstruction will be analyzed and methods of detecting time variations in static reconstructions suggested. Geometric distortion is the most serious type of error caused by time variations in the ionosphere, since geometric distortion produces features that resemble valid ionospheric structures. If the reconstruction algorithm uses relatively strong a priori information, then analysis of the residual and number of negative pixels can be used to detect whether an image is corrupted by movement of the source.; This thesis will present the first algorithm for time varying CIT. The new algorithm is a noniterative algorithm which extends the usual two dimensional reconstruction along a third axis representing time. The new algorithm will be used to analyze the configuration of time varying CIT systems.