Study On Refined Model And Iterative Solving Methods Of GNSS Water Vapor Tomography

Water vapor is an important component of the lower atmosphere and plays a very important role in maintaining the environmental temperature suitable for life on Earth,the hydrological cycle,and many other aspects.Moreover,the process of strong water vapor phase change directly affect the vertical stability of atmospheric structure,which may form destructive/extreme weather events.Therefore,accurate monitoring of the distribution and spatio-temporal variation of atmospheric water vapor is of great significance for the prediction and warning of short-coming extreme weather events and the study of the mechanism of long-term climate change.At present,ground-based GNSS water vapor detection technology is one of the most potential means.Using the data from the network of ground stations can not only obtain continuous,low-cost,high-precision,high-spatial-temporal resolution and allweather-availability two-dimensional water vapor information,but also can reconstruct the three-dimensional spatial and temporal distribution of water vapor by GNSS tropospheric tomography technology.In this thesis,the principles and methods of precipitable water vapor retrieval using ground-based GNSS and obtaining three-dimensional water vapor density/wet refractivity by GNSS water vapor tomography are described in detail.On this basis,the key techniques such as the construction of tomographic model and the solution of the tomographic equations are studied and some useful results are obtained.The main research work and contributions of this thesis are summarized as follows:1.The iterative inversion algorithm for the tomographic equations is examined.(1)A two-step projected iterative algorithm is proposed.In this method,a hypothesis convex set of unknowns is constructed by using prior water vapor data to constrain the reconstruction in the iterations.Because the additional constraints that incorporate a priori knowledge of the solution are added to the reconstruction,it is typically leading to a better tomographic result.Moreover,this method also includes a step of preprocessing for the initial values to reduce the adverse effect of the empirical initial values on the results.The experimental results in Hong Kong network show that compared to the classical inversion methods,the proposed method can reduce the reconstruction errors and improve the accuracy of the tomographic results effectively.The accuracy of the reconstructed water vapor field is improved by about 10% and 15%when radiosonde and ECMWF data were used as the reference,respectively.(2)The convergence property of the iterative method,i.e.,semi-convergence,is studied and discussed in the tomographic equations.Then,the strategy for selecting the relaxation parameters is given based on the semi-convergence property.(3)A new iterative termination criterion(2D-NCP)is proposed,in which the observations are grouped and processed according to their different altitude angles and stations.Experimental results illustrate that the criterion is advantageous in terms of both computational efficiency,compared to the conventional termination criteria,and avoidance of complex work on the determination of the threshold for the termination of the iterative process.2.The method of incorporating the signals leaving the tomographic region laterally are studied.A refined extended tomographic model with combining virtual signals is proposed to improve the distribution of observations and enhance the performance of tomographic solutions.The new model was established by adding some auxiliary voxels around the original tomographic region,which is inverted cone-shaped.The virtual observation signals that cross from the side-faces of the original model and penetrate from the top boundary of the extended region(mainly at low elevations)were constructed and introduced into the tomographic equations.The SWD of the virtual rays were computed using the tropospheric parameters from GNSS data processing.The tomographic results under different weather conditions are verified using radiosonde and ECMWF data as the reference.One can see that the proposed method can obtain better water vapor distribution than the standard tomographic model,with the accuracy improvements of 12% and 22%,respectively.There are 29 figures,6 tables,142 references in this thesis.