A Technology Study of Fof2 Forecasting During the Ionospheric Disturbance

Abstract The ionosphere is an especially complicated and nonlinear system. For nonlinear mapping between input and output of network, the artificial neural back propagation (BP) network is adopted to forecast f o F 2 , which depends on solar activities, geographical positions and the season, etc. Ionospheric sounding data at Haikou and Changchun, China, is used in this study, corresponding respectively the lower and middle latitudes. Ionospheric disturbance events during years 1994 and 2001 are referred as of lower and higher solar activity conditions. Compared with IRI (International Reference Ionosphere), results show significant advantage of ANN (Artificial Neural Network) method in forecasting ionospheric disturbances. In addition, the residual error model is established based on forecasting and observations during the years 2000 and 2001 at Haikou, China, which is essentially useful in many cases.