Ionospheric irregularities are a regular occurrence at the equatorial latitude during the postsunset hours especially during high-solar activity. These irregularities could pose serious challenges to satellite-based navigation and positioning applications by causing fading and degradation of transionospheric signals passing through these irregularities. We have investigated large-scale ionospheric irregularity occurrence at Ilorin, Nigeria (latitude = 8.48°N, longitude = 4.67°W, dip = 4.1°S), a station located within the equatorial region in the African sector. The index used in this study is the rate of change of total electron content (rate of change) derived from 30 s receiver-independent exchange data obtained using a dual frequency GPS receiver (i.e., NovAtel GPStation-2). The study covers a period of 4 years (2009–2012). The results obtained showed that large-scale irregularities occur between March and November and are more pronounced between 1900 LT and 2400 LT. The irregularities were observed to show two peaks: one in March and the other in September. Solar activity trend was also observed. The irregularity level around the peaks seems to increase with solar activity. Although the study covered a period of 4 years, the period could be regarded as the increasing phase of the solar cycle 24.

The limitation of ionospheric models in describing short-term ionospheric events has led to the development of data assimilative models e.g. the International Reference Ionosphere extended to Plasmasphere (IRI-Plas) model. This paper compares the IRI-Plas derived total electron content (TEC), the peak height (hmF2) and critical frequency (foF2) of the F2-layer with those obtained from Global Positioning System (GPS) receiver's and Digisonde Precision Sounder (DPS-4) measurements over Ilorin (Geog. Lat. 8.50^oN; Long. 4.50^oE, dip: - 7.9^o) during geomagnetic storm days. The model estimation was done by assimilation of Ionosonde foF2 and TEC derived from GPS (GPS-TEC) and Global Ionospheric Map (GIM-TEC) into the model code. In order to study the effect of data assimilation on the model's representation, the "no input" option of the model was used as reference. The result shows that with the exception of the foF2 assimilation mode, all the options generally reproduced TEC quite well for all the storm days considered. Overall, the model adjusted with GPS-TEC gives the best prediction of TEC as it reduced the prediction error of TEC by a multiple of up to three compared to using the GIM-TEC. Also, all the options failed to reproduce the storm induced prominent features in the storm-time features of foF2 and hmF2. In other word, assimilation with the TEC does not generally improve the storm-time predictions of foF2 and hmF2 at the station. Consequently, for storm-time estimation of the F2-layer peak parameters, the 'no input' representation of the model is more valid at this station.