Study On Regional Ionospheric Model Based On Gnss

The ionosphere is an important part of the Sun-Earth space, the existence of the ionosphere makes life on Earth to avoid the direct effect of the sun's ultraviolet radiation and X-rays, but makes radio wave propagation in the ionosphere produce reflection, scattering, refraction and absorption effect, resulting in a serious impact on radio communications, space activities and global positioning system navigation and so on. Therefore, actively monitor and study various phenomena in the ionosphere, which reveals the physical mechanism of the phenomenon, exploring the impact of the ionosphere on the human environment, is an important part of space weather research, and is very important to the development of social economy.This thesis focused on the existing problems of ionospheric model research based on the actual observation data of the ionosphere and the CORS (Continuous Operational Reference System) data of GNSS receiving station, such as the regional ionosphere forecast method and regional ionospheric tomography method. The main work and results are:(1) A single station critical frequency of the F2 layer (foF2) prediction model (called BP-IRI model) is developed based on the neural network and error compensation technology. Using the hourly values of f0F2 within 27 years (1958-1984 years) over Lycksele, Sweden, and the corresponding geomagnetic indices, solar activity index, diurnal variation and seasonal variation, and f0F2 prediction value of IRI2012 model, the prediction capability of the proposed method is verified and analyzed. According to the statistical analysis of average RMSE, the BP-IRI model offers an improvement of 19.1% over the IRI2012 model.(2) A regional VTEC short-term forecasting model (called BP-DPM model) is proposed by combined with the traditional two-dimensional polynomial region ionosphere model (2-DPM) and the CORS data of China's Jiangsu province. Through the analysis of a specific example, the results show that the proposed BP-DPM model gives an average BP-DPM model improvement of 14.0% compared to 2-DPM.(3) A new multiscale ionospheric tomography (MST) technique is developed based on the CORS data of China's Jiangsu province for the tomographic reconstruction of the ionospheric electron density (IED) distribution, which is a simultaneous application of many overlapping single-scale ionospheric tomographies (SSTs) with different pixel sizes. In contrast to conventional SST which parameterizes the model with non-overlapping pixel, the MST allows a decomposition of the ionosphere into a set of SSTs with different pixel sizes; at each model location the IED receive different contributions from all the MST submodels in the process of inversion. The final MST model is a post-inversion superposition of all submodels and the final MST solution is a superposition of solutions of all submodels. In comparison with the conventional SST models, superior results of the MST are observed in a numerical simulation. A careful validation of the reliability and superiority of MST is made. According to the statistical analysis of average root mean square error, the result shows that the proposed model offers an improvement of 20.8%,22.2%, and 77.7% over the conventional CSST (SST model with CART algorithm), SST, and IRI2012 models, respectively.(4) A new variable pixel height computerized ionospheric tomography (VHCIT) technique is developed based on the CORS data of China's Jiangsu province for the tomographic reconstruction of the ionospheric electron density (IED) distribution, in which the height of each pixel scale is considered as a variable, which is simultaneous solved together with the unknown electron density values based on the traditional ionospheric tomography model (SST). As a result, this is equivalent to the ionospheric structure of inversion region were second or multiple partitioned after the model inversion ended, and then reconstructed the ionospheric electron density (IED) distribution. According to the GNSS observation data, a careful validation of the reliability and superiority of VHCIT is made. Compared to other traditional ionospheric tomography models, more accurate and realistic electron density distribution is reconstructed by VHCIT model. According to the statistical analysis of average root mean square error, the result shows that the proposed model offers an improvement of 8.6%,14.8%, and 75.3% over the conventional CSST, SST, and IRI2012 models, respectively.