Analysis and Modeling of Temporal Variation Properties for LF Ground-Wave Propagation Delay

In order to further improve the predictive accuracy of low-frequency (LF) ground-wave propagation delay, it is necessary to establish a proper model that can take complex temporal variation properties of propagation delay into account. In this letter, we first analyze the relationship between measured propagation delay and five typical meteorological factors. It can be found that propagation delay has strong correlation with meteorological factors, for example, there is a positive linear correlation between propagation delay and temperature. Then, a theoretical model of propagation delay is established by using a backward propagation neural network (BPNN), and its accuracy is validated by the good agreement between predicted results with measured values. Finally, a detailed quantitative comparison shows that on one hand, the more meteorological factors are considered, the more accurate the predictive model is; on the other hand, the BPNN provides a very convenient approach to handle a large number of factors.