The ionospheric polarisation of very low frequency radio waves.

摘要

The lowest regions of the ionosphere are conveniently studied by means of reflection techniques using very low frequency radiowaves (10-30 kHz.).Most observations of this kind have measured only the cross polarised component of the reflected signal yielding the so-called conversion coefficient, R1. In this thesis an attempt is made to measure the in-plane signal, R, in addition to R1. From these two parameters the polarisation of the reflected waves may be determined. The results are interpreted in terms of the D-region electron density profiles by modelling techniques based on the full-wave analysis. Data collection for a period of one year are presented and discussed. A detailed account is given of the construction of equipment suitable for measurement of the normal polarised component of the ionospherically reflected radiowave, using the a modification of the 'loop and aerial' system. Due consideration is given to the experimental difficulties. It is established that the stability, both in phase and amplitude, is an extremely important factor in the success of the experiment. Methods are discussed of ensuring that no ground wave contamination exists, and details are given of a computer program to do this. A large section of the disertation is devoted to a detailed descrimption of the development of the data logging systems as these were responsible, to a large extent, for the success of the experimental programme. Of particular importance is the description of the on-line computing facility, using an advanced programmable calculator, with magnetic tape facility. The experimental data collected over one year is presented, and both seasonal and diurnal variations are presented. The related electron density profiles, derived by modelling the four measured paramenters, are discussed. Disturbed conditions have been considered for times of solar flare, stratospheric warmings and magnetic storm activity. Again modelling of profiles has been undertaken and a representative set is given. The necessity of measuring both, R and R1 for modelling work has been demonstrated and the errors likely to occur from using R1 alone have been fully discussed.