Research On GNSS Three-dimensional Water Vapor Tomography Model

As a changeable composition in the atmosphere,water vapor plays a very important role in atmospheric energy transport and evolution of weather systems.Now,based on GNSS technology,achieving the information acquisition of tropospheric water vapor and monitoring extreme weather has become one of the hotspots in GNSS meteorology.Based on precise point positioning technology,obtaining slant wet delay(SWD)by using the GNSS water vapor inversion technique,and establishing the joint observation equation.The water vapor reconstruction algorithms are used to reconstruct the spatial distribution of three-dimensional water vapor content over the study area.The main contents and conclusions are as follows:1.The accuracy of GPT,GPT2 and GPT2 w models was compared by using the meteorological measurements from IGS stations.The results show that: GPT2 and GPT2 w model have the same accuracy,which are superior to GPT model;PWV derived from GPT2 and GPT2 w model have the same accuracy that the average bias is 1.2 mm and the average RMS is 1.8 mm,which are better than GPT model.Experiment result provided referential significance for model improvement and application.2.The accuracy of five Tm models were compared by using the ERA-Interim and IGRA data.The results show that: the accuracy of GPT2 w model is the best,the average bias and RMSE values are 0.38 K and 3.62 K respectively;the five models were affected by seasonal factors,and the Tm values estimated by the model were the best in summer precision;RMSE of five models become worse with elevation increasing;The latitude of grids have great effect on the these models.The accuracy of Tm from these models in high latitude areas is poor;PWV derived from five models have the same accuracy that the average RMSE are 1.2-1.4 mm;compared with radiosonde data;the accuracy of GPT2 w model is slightly better than the other four models,witch the RMSE value is 3.79 K.The PWV derived from five models have the same accuracy,and the RMSE are kept between 1.2-1.4mm.3.Comparing the results of the algebraic reconstruction algorithm and the joint algebraic reconstruction algorithm with the results of the radio sounding data in the GNSS three-dimensional water vapor chromatography test over the Hongkong area.the root-mean-square-deviations between ART and SIRT results and radiosonde data are both accurate,the average RMS is 4mm/km.but SIRT algorithm is shorter in solving time,only cost 1/3 of the time of convergence compared with the algebraic reconstruction.Therefore,in better solving the GPS three-dimensional water vapor tomography,the SIRT algorithm is superior to the ART algorithm.4.The IGS prediction ephemeris is used to search for the best boundary of the study area,and the multiple mesh generation schemes are determined.Then using IGS precise ephemeris and station data to carry out the three-dimensional tomography.The results show that the non uniform stratification scheme are better than the uniform stratification.When the size of the horizontal grid is 0.08 degrees,and the accuracy of the computed tomography is the best.