Spatial Interpolation And Tomography Of Ground-based GNSS-derived Water Vapor

Ground-based GNSS water vapor retrieval can provide a new method for tropospheric water vapor detection.This method is all-weather and all-time supporting.Besides,the products come from this way with high accuracy and low cost compared to traditional methods.These products consist of 2D and 3D water vapor distribution,the later came from tomography.Nowadays,the water vapor products from this technology have high temporal and spatial resolution given the extensive construction of CORS stations.The objective of the study is to establish high accuracy weighted mean temperature model,interpolate GNSS-derived PWV,and solve water vapor tomographic issues.The thesis chose Hong Kong as the study area,the aforementioned researches were conducted based on GNSS observations and ground meteorological data provided by Hong Kong CORS Network,and sounding data came from King's Park radiosonde.The main contents of this work are as follows:1.The method for modeling multifactorial weighted mean atmospheric temperature model.This study established two multifactorial models for Hong Kong by stepwise regression and principal component regression.Test results show that the prediction performance of the two models with the RMSE less than 2.1 K are better than Bevis model obviously.2.An investigation of PWV spatial interpolation.At present,the distribution of GNSS stations in CORS network is sparse and uneven,resulting in a low horizontal resolution in GNSS-derived PWV map.To solve the aforementioned problem,this research proposed a new method for PWV interpolation.The proposed method adopted PWV data came from virtual sampling points as input.These virtual sample points originated from the selected DEM data were constructed on the basis of the statistically significant correlation between PWV and geographical location.The advantages of the proposed method were verified by comparing with the conventional method which only takes the original sampling points into consideration.3.Construction and solution of water vapor tomographic equations.In this thesis,we proposed a method to simplify the constructing process of the coefficient matrix for the tomographic equations.Based on the space vector,this method determines whether the intersection point is in the study area by the location of the intersection point.In this study,Kalman filter was used to solve the tomographic equations,and the influence of filtering frequency towards the accuracy of tomographic solutions was analyzed.Test results show that the GNSS moisture profile computed by Kalman filtering can reflect the vertical distribution characteristics of water vapor accurately,and the filtering frequency has no significant effect on the accuracy of the tomographic solutions.