Prediction of Ionospheric TEC Based on ESOA-BP Neural Network

The global total electron content (TEC) of the ionosphere is an important characteristic parameter of the ionosphere, and the realization of ionospheric TEC forecasts is of great value in space science research, as well as in engineering applications such as satellite navigation, ground-based radar, and SAR. This paper applies an improved BP based on Enhanced Seagull Optimization Algorithm (ESOA) to forecast global ionospheric total electron content (TEC) maps with up to 24 hr. The ESOA-BP model is trained and evaluated using Center for Orbit Determination in Europe (CODE) global TEC maps from 2018 to 2022. An analysis of the effectiveness of forecasting solar activity in high and low years, as well as during periods of geomagnetic storms, shows that the correlation coefficient from model TEC concerning the CODE TEC are 0.96 and 0.94 in low solar activity years and high solar activity years, and TEC forecast errors increase during the onset of magnetic storms compared to periods of no storm.