High-precision Time-domain Ionospheric Prediction Model With Physical Parameter Constraints And Its Application In PPP-RTK

Considering the influence of solar activity and geomagnetic change on the ionospheric change,a high-precision ionospheric prediction model(Physical constrained polynomial model,PCPM)is constructed by adding KP index,DST index,f10.7p,R sunspot number and other geomagnetic and solar change indexes as physical parameter constraints.Based on the data from CORS stations in China,the ionospheric prediction is analyzed,and compared with the autoregressive integrated moving average(ARIMA)model.The prediction performance of the two models in different time span,different latitude position and different prediction periods are compared,as well as the effect analysis of the model products on PPP-RTK.The experimental results are as follows:1)Under different time span.On average,The residual of 72.01% results of seasonal ARIMA model are within 1 TECU,95.21% are less than 2 TECU,and 0.63%are more than 3 TECU;82.85% of PCPM prediction residuals are less than 1 TECU,10.85% higher than ARIMA model;95.14% of PCPM data residuals are less than 2TECU,and there is no data with residuals greater than 3 TECU.In the period of ut02-14,ARIMA model has the highest accuracy in the evening and the worst in the afternoon,and PCPM has the best prediction effect in ut04-10(local time afternoon in the study area).The accuracy of PCPM on the first day is better than that of seasonal ARIMA model.With the increase of prediction time span,the accuracy of the two models decreases as a whole.The accuracy of the first three days of seasonal ARIMA model is the lowest(1.5tecu).ARIMA model is better than PCPM in long time span.2)According to the model performance at different latitudes,The average precision of PCPM in mid latitude is 0.81 tecu and 1.40 tecu respectively,which is 15.5% and 23%higher than that of seasonal ARIMA model.PCPM is more feasible in low latitude area,and its regional accuracy is better than seasonal ARIMA model in China.3)Under different forecast periods.The results show that the longer the modeling period is,the higher the accuracy of the model is.The accuracy of the 27-day data forecast results is 0.1-0.3tecu,higher than that of the 10-day data forecast results in ut02-14 time.The overall accuracy of the 27-day forecast results is 0.97 tecu,which is0.14 tecu higher than that of the 10 day forecast results(1.11tecu),with the accuracy increased by 12.61%.In the actural production process,the accuracy of one-day prediction of 10-day data is about 1 TECU,which has met the needs of practical application.4)In the PPP-RTK experiment,the seasonal ARIMA model prediction product can effectively improve the fixed speed of ambiguity.Compared with the TEC constraint in the same period of the previous day,the fixed speed of ambiguity of the model prediction product is increased by 11.5%,25% compared with the fixed speed of the ionosphere free product;the PCPM can improve both the fixed speed of ambiguity and the fixed precision of ambiguity.Compared with the result without ionosphere constraint,the accuracy of e-direction and u-direction of seasonal ARIMA model is increased by 16.7%,61.0% and 6.9%,respectively.