GNSS Radio Occultation (GNSS-RO) is an opportunistic Earth sensing technique where GNSS signals passing through the atmosphere are received in low Earth orbit and processed to extract meteorological parameters. As signals are received along an orbit, the measured Doppler shift is transformed to a bending angle profile (commonly referred to as bending angle retrieval), which, in turn, is inverted to a refractivity profile. Thanks to its high vertical resolution and SI traceability, GNSS-RO is an important complement to other Earth sensing endeavors. In the lower troposphere, GNSS-RO measurements often get degraded and biased due to sharp refractive gradients and other complex structures. The main objective of this thesis is to explore contemporary retrieval methods such as phase matching and full spectrum inversion to improve their performance in these conditions. To avoid the bias caused by the standard inversion, we attempt to derive additional information from the amplitude output of the examined retrieval operators. While simulations indicate that such information could be found, it is not immediately straightforward how to achieve this with real measurements. The approach chosen is to examine reflected signal components and their effect on the amplitude output.

Radio Occultation (RO) is a remote sensing technique which uses Global Navigation Satellite System (GNSS) signals tracked by a Low-Earth Orbit (LEO) satellite to sound the earth's atmosphere both in low (troposphere, stratosphere) and high (ionosphere) altitudes. GNSS-RO provides global coverage and SI traceable measurements of atmospheric data with high-vertical resolution. Refractivity, dry temperature, pressure and water vapour profiles retrieved from RO measurements have a relevant contribution in Numerical Weather Prediction (NWP) systems and in climate-monitoring. Due to the partial propagation through the ionosphere, a systematic bias is added to the lower atmospheric data product. Most of this contribution is removed by a linear combination of data for two frequencies. In climatology studies, one can apply a second-order correction - so called κ-correction - which relies on a priori information on the conditions in the ionosphere. However, both approaches do not remove high-order terms in the error due to horizontal gradient and earth's geomagnetic fields. The remaining residual ionospheric error (RIE) and its systematic bias in RO atmospheric data is a well-known issue and its mitigation is an open research topic. In this licentiate dissertation, the residual ionospheric error after the standard correction is evaluated with computational simulations using a wave optics propagator (WOP). Multiple Phase Screen (MPS) method is used to simulate occultation events in different ionospheric scenarios, e.g. quiet and disturbed conditions. Electron density profiles (EDP) assumed in simulations are either defined by analytical equations or measurements. The disturbed cases are modelled as small-scale irregularities within F-region in two different ways: as sinusoidal fluctuations; and by using a more complex approach, where the irregularities follow a single-slope power-law that yields moderate to strong scintillation in the signal amplitude. Possible errors in MPS simulations assuming long segment of orbit and ionosphere are also evaluated. The results obtained with the sinusoidal disturbances show minor influence in the RIE after the standard correction, with the major part of the error due to the F-region peak. The implementation of the single-slope power-law is validated and the fluctuations obtained in simulation show good agreement to the ones observed in RO measurements. Finally, an alternative to overcome limitations in MPS simulations considering occultations with long segment of orbit and ionosphere is introduced and validated. The small-scale irregularities modelled in F-region with the power-law can be added in simulations of a large dataset subjected to κ-correction, in order to evaluate the RIE bending angle and the consequences in atmospheric parameters, e.g. temperature. / NRPF-3, Rymdstyrelsen, 241/15