Modeling Ionospheric FoF2 In China And Its Research On Application

The ionosphere is about 50 kilometers above the earth's surface,which is filled with the neutral atmosphere and plasma,and it is also the junction between the deep crust,the earth's surface,the middle atmosphere,the magnetosphere and the interplanetary space.The study of seismic-ionospheric effect began in the 1960s.When American scientists studied the Alaska M8.5 earthquake in 1964,they first found the phenomenon that the disturbance of plasma density before this earthquake.Since then,many scholars have been engaged in the study of seismic-ionospheric effects.Along with the Chinese seismo electromagnetic satellite "ZH-1" successfully launched in Jiuquan,it has opened up a new way to solve the earthquake prediction which is the scientific problem all over the world.In general,a detailed analysis for the background changes of ionospheric parameters is essential before the ionospheric parameter information is used to identify seismic abnormal signals.As the distribution of ionospheric F2 layer critical frequency(foF2)stations in China is sparse,in order to improve the accuracy of foF2 prediction model,in this study the totel electron content(TEC)prediction model of China was built first by using artificial neural network,and then the foF2 model is inversed by the coupling relationship between foF2 and TEC.This paper illustrates the application of neural networks(NN)in developing a regional prediction model for the ionospheric total electron content(TEC)over China.To avoid the 'local minimum' effect caused by the traditional NN-based model,genetic algorithm(GA)is utilized to optimize the initial weights of NN.The NN in this study has 19 input parameters,which are concerned with the ionospheric diurnal variations,seasonal information,solar cycle,geomagnetic activities,geographic coordinates,and declination.The output parameters are the daily hourly values of vertical total electron content(VTEC)detected from 43 permanent GPS(Global Positioning System)stations in China.The datasets during 2012-2014 are used to train the network,and the datasets of 2015 are selected as the test data to verify the model precision.Prediction results of the genetic algorithm-based neural network(GA-NN)model,back propagation-based neural network(BP-NN)model,and International Reference Ionosphere 2012(IRI 2012)model are put together to compare with the observed TEC data from 12 GPS stations in China.According to the numerical analysis,the root-mean-square-error(RMSE)band obtained from GA-NN is 5.2140-8.4756 TECU,percent deviation(PD)is 8.78-12.30%,and the range of correlation coefficient is 0.9583-0.8069,while those form the BP-NN and IRI2012 models are 6.2962-12.1468 TECU,10.17-14.16%,0.7192-0.9348,and 6.5513-11.7937 TECU,11.01-14.07%,0.7292-0.9129,respectively.Then combined with the comparisons between the diurnal variations of TEC values,the performance of GA-NN model is still much better than BP-NN and IRI 2012 models.Subsequently,the seasonal and local variation characteristics are also validated based on the GA-NN model.The results indicate that the regional TEC prediction model developed by GA-NN is very promising for ionospheric studies.Since the coupling relation between foF2 and TEC is not a fixed linear relation,the coupling relation between TEC and foF2 is searched by GA-NN method again,by which the foF2 model of China region is constructed.By comparing the prediction results between the GA-NN model and the IRI2012 model via RMSE,PD and p,the results show that the prediction of ionospheric parameter model by GA-NN technology has great prospects in ionospheric research.Based on the data of GPS TEC,taking Myanmar M7.2 earthquake on April 13th,2016 as an example,this paper analyzes the spatial and temporal distribution of seismo-ionospheric anomalies in China by using upper and lower thresholds to extract disturbance.The result shows 2 to 4 days before this earthquake,TEC disturbance in China was seismo-ionospheric precursor;especially,the ionospheric disturbance on April 11th(two days before earthquake)was the most affected by seismic activity,which the farthest influence range of disturbance could be about 2816km;meanwhile,on April 11th,the most obvious abnormal area was in the north of the epicenter direction(in South China).Through extracting the TEC data from Jet Propulsion Laboratory(JPL)on the longitude chains of the most obvious abnormal area,it was found that the north EIA had been moved northwards.Then the possible physical mechanism of this abnormal phenomenon had been explained by using the theory of electrostatic field.