Empirical Modeling of Ionospheric Current Using Empirical Orthogonal Function Analysis and Artificial Neural Network

Given the potential importance of solar quiet (Sq) ionospheric current in geomagnetic field modeling, it is vital to obtain accurate parameters characterizing its variations, particularly spatial and temporal variations. In this paper, we derived Sq function based on spherical harmonic analysis (SHA) technique using a 14-year (2006–2019) record over American sector. The empirical orthogonal (EOF) was then applied deduce variations current. It observed that first EOF mode dominated by activity, while second third modes exhibit annual semiannual respectively. Also, artificial neural network (ANN) model constructed validate predictions. While intensity predicted ANN underestimated 2.83%, underpredicted 1.92% relative observation. root mean square error (RMSE) 0.64 kA. This RMSE about 79% smaller than model. addition, both models capture variation total (Jtotal) with respect activity. principle, had an optimal performance at nearly 98% accuracy, exhibiting almost same degree which appears be reference point for conditions when looking space weather applications.