Equatorial F-region plasma density estimation with incoherent scatter radar using a transverse-mode differential-phase method

This dissertation presents a novel data acquisition and analysis method for the Jicamarca incoherent scatter radar to measure high-precision drifts and ionospheric density simultaneously at F-region heights. Since high-precision drift measurements favor radar return signals with the narrowest possible frequency spectra, Jicamarca drifts observations are conducted using the linear-polarized transverse radar beams. Transverse-beam returns are collected using an orthogonal pair of linear-polarized antennas, and the average power as well as phase difference of the antenna outputs are fitted to appropriate data models developed based on the incoherent scatter theory and the magneto-ionic theory. The crude differential-phase model when is characterized in terms of straight line fields is applied to the January 2000 data. The most complete differential-phase model, which takes into account the misaligned angle between the dipole axes and geomagnetic northeast and southeast directions, as well as the radar beam width and variation of magnetic fields, is applied to the January 2000 data and June 2002 data. We present and compare the inversion results obtained with different versions of the data models and conclude that the geometrical details have only a minor impact on the inversion. We also find that the differential-phase method works better for the 15-min integrated January 2000 data than 5-min integrated June 2002 data since the former has the bigger densities, larger SNR of the backscattered signals, and more usable phase data. Our inversion results show reasonable agreement with the ionosonde data.; The full correlation method is formulated and applied to the June 2002 data. Compared to the differential-phase method, this method is different in the sense that it utilizes the real and imaginary parts of the cross-correlation of orthogonal antenna outputs at the high altitudes where SNR is low and the off-diagonal elements of the covariance matrix of measurement errors can be ignored. An advantage of this is that the cross-correlation data have Gaussian distributed errors even when the SNR is weak, whereas under the same conditions the phase errors are no longer Gaussian. The determination of noise levels has been improved with the full correlation method.