Space weather and its effect on technological systems are important for scientific research. Developing an understanding of the behaviour, sources and effects of ionospheric events form a basis for improving space weather prediction. This thesis attempts to use co-located radars and instruments for the analysis of ionospheric events over South Africa. The HF Doppler radar, ionosonde, Global Positioning System (GPS) and GPS ionospheric scintillation monitor (GISTM) receivers are co-located in Hermanus (34.4°S, 19.2°E), one of the observatories for the space science directorate of the South African National Space Agency (SANSA). Data was obtained from these radars and instruments and analysed for ionospheric events. Only the Hermanus station was selected for this analysis, because it is currently the only South African station that hosts all the mentioned radars and instruments. Ionospheric events identified include wave-like structures, Doppler spread, sudden frequency deviations and ionospheric oscillations associated with geomagnetic pulsations. For the purpose of this work, ionospheric events are defined as any unusual structures observed on the received signal and inferred from observations made by the HF Doppler radar. They were identified by visual inspection of the Doppler shift spectrograms. The magnitude and nature of the events vary, depending on their source and were observed by all, some or one instrument. This study suggests that the inclusion of a wider data coverage and more stations in South Africa merit consideration, especially since plans are underway to host a co-located radar network similar to that in Hermanus at at least three additional observatory sites in South Africa. This study lays a foundation for multi-station co-located radar and instrument observation and analysis of ionospheric events which should enhance the accuracy of space weather and HF communication prediction.

In this thesis an ionospheric tomographic algorithm called Multi-Instrument Data Anal- ysis System (MIDAS) is used to reconstruct electron density profiles using the Global Positioning System (GPS) data recorded from 53 GPS receivers over the South African region. MIDAS, developed by the Invert group at the University of Bath in the UK, is an inversion algorithm that produces a time dependent 3D image of the electron density of the ionosphere. GPS receivers record the time delay and phase advance of the trans- ionospheric GPS signals that traverse through the ionosphere from which the ionospheric parameter called Total Electron Content (TEC) can be computed. TEC, the line integral of the electron density along the satellite-receiver signal path, is ingested by ionospheric tomographic algorithms such as MIDAS to produce a time dependent 3D electron density profile. In order to validate electron density profiles from MIDAS, MIDAS derived NmF2 values were compared with ionosonde derived NmF2 values extracted from their respective 1D electron density profiles at 15 minute intervals for all four South African ionosonde stations (Grahamstown, Hermanus, Louisvale, and Madimbo). MIDAS 2D images of the electron density showed good diurnal and seasonal patterns; where a comparison of the 2D images at 12h00 UT for all the validation days exhibited maximum electron concentration during the autumn and summer and a minimum during the winter. A root mean square error (rmse) value as small as 0.88x 10¹¹[el=m³] was calculated for the Louisvale ionosonde station during the winter season and a maximum rmse value of 1.92x 10¹¹[el=m³] was ob- tained during the autumn season. The r² values were the least during the autumn and relatively large during summer and winter; similarly the rmse values were found to be a maximum during the autumn and a minimum during the winter indicating that MIDAS performs better during the winter than during the autumn and spring seasons. It is also observed that MIDAS performs better at Louisvale and Madimbo than at Grahamstown and Hermanus. In conclusion, the MIDAS reconstruction has showed good agreement with the ionosonde measurements; therefore, MIDAS can be considered a useful tool to study the ionosphere over the South African region.

Ionospheric storms are an extreme form of space weather phenomena which affect space- and ground-based technological systems. Extreme solar activity may give rise to Coronal Mass Ejections (CME) and solar flares that may result in ionospheric storms. This thesis reports on a statistical analysis of the ionospheric response over the ionosonde stations Grahamstown (33.3◦S, 26.5◦E) and Madimbo (22.4◦S,30.9◦E), South Africa, during geomagnetic storm conditions which occurred during the period 1996 - 2011. Total Electron Content (TEC) derived from Global Positioning System (GPS) data by a dual Frequency receiver and an ionosonde at Grahamstown, was analysed for the storms that occurred during the period 2006 - 2011. A comprehensive analysis of the critical frequency of the F2 layer (foF2) and TEC was done. To identify the geomagnetically disturbed conditions the Disturbance storm time (Dst) index with a storm criteria of Dst ≤ −50 nT was used. The ionospheric disturbances were categorized into three responses, namely single disturbance, double disturbance and not significant (NS) ionospheric storms. Single disturbance ionospheric storms refer to positive (P) and negative (N) ionospheric storms observed separately, while double disturbance storms refer to negative and positive ionospheric storms observed during the same storm period. The statistics show the impact of geomagnetic storms on the ionosphere and indicate that negative ionospheric effects follow the solar cycle. In general, only a few ionospheric storms (0.11%) were observed during solar minimum. Positive ionospheric storms occurred most frequently (47.54%) during the declining phase of solar cycle 23. Seasonally, negative ionospheric storms occurred mostly during the summer (63.24%), while positive ionospheric storms occurred frequently during the winter (53.62%). An important finding is that only negative ionospheric storms were observed during great geomagnetic storm activity (Dst ≤ −350 nT). For periods when both ionosonde and GPS was available, the two data sets indicated similar ionospheric responses. Hence, GPS data can be used to effectively identify the ionospheric response in the absence of ionosonde data.