Performance Analysis of a Regional Ionospheric Model Over Nigeria Within Equatorial Ionospheric Anomaly Region.

The development of regional ionospheric models have become increasingly important due to regional and local ionospheric processes that dominate regional responses to ionospheric drivers. The equatorial ionosphere is noted to be complex and very dynamic. Nigeria is located within the complex equatorial ionization anomaly region. The Centre for Atmospheric Research Neural Network (CAR-NN) used in this work was developed by the method of artificial neural networks having 8 input layer neurons, 11 hidden layer neurons, and 1 output layer neuron. The observed Vertical Total Electron Content (VTEC) values retrieved from ten (10) GPS stations well-distributed over Nigeria, within equatorial ionospheric anomaly region, were analyzed and compared with corresponding values obtained from the International Reference Ionosphere (IRI) model, NeQuick and CAR-NN models. CAR-NN used in this work is considered as a regional model, while the IRI and NeQuick are well-known global ionospheric models. We evaluated TEC from the three models at hourly levels for all the days of the year 2012. TEC obtained from the ten stations under consideration was also evaluated for all the days of the year. A comparison of VTEC predictions from the model used in this work with those from the NeQuick and the IRI shows that CAR-NN produced better estimates of TEC than the IRI and NeQuick at all locations and time. The regional CAR -NN model is better in terms of the prediction error after we computed the RMSE for the three models.

Keywords: Ionosphere, Models, GPS