Research On The Distribution Of Water Vapor By Combining Ground-based And Space-based GNSS

Global climate change and frequent extreme weather events threaten human survival and development.High-precision,high temporal and spatial resolution of atmospheric vapor distribution is an important source of information for monitoring and forecasting global climate change and severe weather.As a new technology for water vapor detection using ground-based GNSS and space-based GNSS,it can effectively compensate for the shortcomings of the traditional water vapor detection technology.At the same time,it also has the advantages of quasi-real-time,all-weather,and no-human interference,and the observation data have long-term stability.In addition,the continuous improvement of the new generation of GNSS satellite navigation system and the implementation of the second-generation COSMIC radio occultation detection technology have laid a solid foundation for the development the detection of atmospheric water vapor by combining ground-based GNSS and space-based GNSS.This dissertation focuses on the improvement of GNSS detection of water vapor spatial and temporal distribution and deeply studies the theories and methods of GNSS tomographic detection assisted by multi-source atmospheric observation data.Firstly,the thesis studies the theory and method of integrating wireless radiosonde data and GNSS wireless occultation data to assist ground-based GNSS to estimate the slant-lath water vapor constent(SWV).Then,the error characteristics of GNSS rado occultat'ion detection data are analyzed,and the "non-difference" processing method is used to extract the additional phase delay of GNSS and applied to the inversion of GNSS wireless occultation to improve the detection accuracy of occultation low-troposphere;The accuracy of the zenith dry term delay model is evaluated,wireless sounding and wireless occultation historical observations are utilized to develop a regional zenith dry term delay model to improve the estimation quality of zenith dry term delays;aiming at the problem that the classical atmospheric weighted average temperature model has large errors in some areas,an atmospheric weighted average temperature model that takes into account ground temperature,time and elevation multi-factor variables is established;the concept of the top layer of the water vapor layer and the top layer of the water vapor dense layer is defined and the method for measuring the height of the two variables is determined.Based on this,a new method of dividing the tomographic grid is proposed.Finally,the tomographic inversion is conducted by collecting observation data from February 2014 in Hong Kong.The main research contents and relevant conclusions of this paper are as follows:1.The "non-difference" method is proposed to estimate the GNSS wireless occultation additional phase delay.The additional phase delay extracted from the GNSS wireless occultation carrier phase is the basis for inversion of the entire occultation data,and the processing accuracy directly affects the quality of the final atmospheric product.Compared with the "double-difference" method and "single-difference" method processing methods,the "non-difference" method does not need to introduce an auxiliary station to eliminate the effect of receiver clock error.Therefore,the extra phase delay extracted by the"non-difference" method reduces the noise level and also weakens the influence of the atmospheric multipath effect in the low troposphere.In addition,we also compared the relative refractivity bias between ND-derived and European Centre for Medium-Range Weather Forecasts(ECNMWF)analyses data of globally distributed 200 COSMIC occultation events on December 12,2013,and the results indicate that the non-differencing-derived refractivity was better than that of single difference processing strategy below 10km.2.To establish a high accuracy regional zenith troposphere dry term delay model,first,an integral method to estimate the zenith troposphere delay is proposed,which takes into account the variation of the refractive index and pressure on the adjacent height layer.The new integral model can estimate high-precision zenith tropospheric delay values from high vertical resolution meteorological products.Regarding the zenith tropospheric dry delay estimated from the radiosonde data and radio occultation data using the new integral model as a true value,the accuracy of three classical zenith tropospheric dry term delay models(Saastamoinen,Hopfield,and Black)is evaluated.Finally,the regional Saastamoinen model is calibrated using radiosonde and radio occultation history data.Compared to the ZHDs which are obtained by Saastamoinen model using these radiosonde and radio occultation data,the accuacy of new model is improved by 44.6%and 36.8%.3.To develop an atmospheric weighted average temperature(Tm)model that takes into account multivariate factors,firstly,based on the law of approximately index changes in vapor pressure with elevation changes and the law of approximately linear changes in atmospheric temperature with elevation changes in adjacent height layers,we propose a new integral model for accurately calculating atmospheric weighted average temperature based on high vertical resolution meteorological data.Then,based on the radiosonde and wireless occultation data,the atmospheric weighted average temperature calculated by the new integral model is used as the background field,an atmospheric weighted average temperature model considering the surface temperature,time and elevation factors is developed.In addition,the COSMIC profiles collocated with Hong Kong and the 45004th radiosonde station collected from January 1,2016 to October 31,2016 are used to evaluate the accuracy of the new Tm model.Using the radiosonde-derived Tm and COSMIC-derived T.as references,compared with the Bevis model,the statistical results indicate that the Tm accuracy is improved by 45.4%and 34.75%,respectively.4.Further improvement of several key issues in ground-based GNSS tomography.For the first time,we introduce the Gaussian index model as a vertical constraint into the ground-based GNSS tomography technique.The concept of the top layer of the water vapor layer and the top layer of the water vapor dense layer and the method for determining the height are proposed,and the top layer of the vapor layer is regarded as the upper boundary layer of tomographic inversion.Based on these two elevations,a new tomographic grid dividing method in the vertical direction is presented.Select the measured data from 28 days in Hong Kong from February 1,2014 to February 28,2014 for inversion.Regardng the radiosonde data and occultation products in the same area and time as true values,through statistics and analysis,we can see that the deviation RMS of the water vapor density and the sounding products obtained by the optimized tomographic technique above 3.75 km and below 3.75 km are 2.52 g/m3 and 0.86 g/m3,respectively.Compared with the traditional chromatographic results,they increase by 15%and 12%respectively.The deviation RMS of the water vapor density and the wireless occultation products obtained by the optimized tomographic technique above 3.75 km and below 3.75 km are 1.24 g/m3 and 0.72 g/m3,respectively.Compared with the traditional chromatographic results,they increase by 15%and 19% respectively.