Prediction of f2-layer height of the peak electron density (hmf2) over the Southern Africa region using artificial neural network

The ionospheric F2 layer is the essential layer in the propagation of high-frequency radio waves, and height of the peak electron density (hmF2) is one of the important parameter. However, ionosondes are not installed at every location on earth to allow for global measurements of hmF2, especially within the southern African region. This study therefore focuses on developing a regional model for predicting the hmF2 using Artificial Neural Network techniques. In this study, prediction model was developed using year, day of the year, time in 30 minutes intervals, Sunspot Number (SSN), Solar flux at 10.7 cm (F10.7), geomagnetic activity (Kp) and Averaged planetary Index (Ap), longitude, latitude and critical frequency (foF2) as the input. The hmF2 prediction of 2009 and 2014 obtained from proposed model its results during summer and winter was compared with AMTB-2013 and SHU-2015, which are recently parameters of the IRI-2020 model. The result showed that in both 2009 and 2014 the year of low solar activity and high solar activity respectively, ANN over perform other models with minimum RMSE and PRMSE values followed by SHU-2015 and finally AMTB-2013. Therefore, it can be concluded that the architecture and learning efficiency of ANN proposed model are as good as SHU-2015 with slightly difference between them. Although all models need some improvement to increase their accuracy.