An assessment of empirical models for the prediction of the transionospheric propagation delay of radio signals
The objective of this thesis research was to test several empirical models of the ionosphere to see which, if any, might be a better predictor of the ionosphere's total electron content (TEC) (and therefore ionospheric delays) than the GPS single-frequency Broadcast model. A total of four models were tested, namely: (1) The Bent Model; (2) The Ionospheric Conductivity and Electron Density (ICEED) profile model; (3) The 1986 International Reference Ionosphere (IR186); and (4) The GPS single-frequency Broadcast model. Each model was adapted to enable multiple epoch/location predictions of TEC. Model testing was broken down into two distinct stages: (1) The first three models were tested against a 48 station months of Faraday rotation measurements of TEC, from a total of five North American sites and one European site, recorded during three different levels of solar activity from the previous cycle; and (2) All four models were then compared with ionospheric delays recovered from dual-frequency GPS data recorded at two Canadian stations during a period disturbed ionospheric behaviour in February 1991. Comparisons with the Faraday rotation data revealed that the Bent and IR186 models were the best. Comparisons with the GPS data showed that the Broadcast and IR186 models to be the best - the Broadcast model was able to account for approximately 70 to 90% of the daytime ionospheric delay and 60 to 70% of the night-time delay. Based upon the findings of this research some strategies for further related work are suggested.