利用人工神经网络建立我国南海地区电离层模式

AN IONOSPHERIC MODEL IN THE SOUTH CHINA SEAREGION MADE BY ARTIFICIAL NEURAL NETWORK

为了解决我国南海地区短波无线电通信信道的选择问题，本文建立了该区域一种新的电离层模式．模式中使用了人工神经网络技术，神经网络选用了５个预报变量，并用南海地区的６个电离层站的观测资料训练网络．结果表明：这个模式能够反映我国低纬电离层随太阳活动水平、经纬度和时间的变化，它适用于我国南海和东南海地区高频通信的频率预报．

In this paper, a regional ionosphere model in the South China Sea region is builton the basis of artificial neural network. In this model, five variables which stand forthe yearly and monthly mean value of F10.7, longitude, latitude and month areselected. Monthly median of ionosphere critical frequency with the length of about100 years observed at six ionospheric observatories around the region are used tOtrain the netWork. Comparison of predictions with observations shows that the modelcan characterize the variation of the critical frequency in thes regional ionosphere withtime, location and the solar activity level. It is suitable for the prediction of HFcommunication channels in the South China Sea.