A Research On Ground-based GNSS Atmospheric Precipitable Water Vapor And Tropospheric Delay Correction Model Without Meteorological Parameters

Water vapor is an important component of the atmospheric environment and a key indicator of extreme weather and rainfall events.In order to realize the monitoring and early warning of extreme weather and rainfall events,it is necessary to obtain water vapor information with a high accuracy and high spatiotemporal resolution.The remote sensing of precipitable water vapor(PWV)using ground-based GNSS technique is one of the most important research topics in GNSS meteorology.This is due to the fact that the method can overcome the shortcomings of most conventional detection methods apart from the advantages being real-time,continuous and stable with a high precision and high spatiotemporal resolution.In this study,estimating PWV from ground-based GNSS is taken as my primary research focus,the key parameters in the inversion process and tropospheric delay correction model without Meteorological Parameters are studied and analyzed.The main research tasks are listed as follows:(1)The optimization scheme of regional zenith tropospheric delay determination using Hong Kong continuous operation reference station(CORS)network was discussed.By analyzing the influencing factors(Type of satellite ephemeris and Calculating method,number and type of auxiliary stations outside the network,elevation Angle)of the regional zenith troposphere delay solutions,the optimization scheme is put forward scientifically as a meaningful reference for future studies.(2)The Zenith Tropospheric Delay(ZTD)in Asia for a period of 10 years(2007-2017)was investigated through incorporating the Global Pressure and Temperature 2/ Global Pressure and Temperature 2w(GPT2/2w)models with the Saastamoinen model(represented by GPT2 S,GPT2w-1S,and GPT2w-5S respectively).The performance of the integrated model was then assessed in terms of accuracy and spatiotemporal distribution.The back propagation(BP)neural network technique was used to improve the GPT2w-1S model and achieved good results,system bias reduced by 63%.(3)Various weighted average temperature models were introduced.The Hong Kong radiosounde station was used as the experimental station to compare and analyze the accuracy of each model.In terms of the model based on ground meteorological parameters,the accuracy of the single-factor,double-factor and three-factor model was analyzed,and a seasonal two-factor model was proposed,with obvious improvement in accuracy(especially in summer,system bias reduced by 52%).In terms of the model without meteorological parameters,the GPT2 w model was introduced and improved by using the BP neural network technique.The improvement of the new model is significant,system bias reduced by 75%.(4)The PWV automatic calculation platform is established by using MATLAB software,which has good accuracy.The correlation analysis of the key parameters in the PWV calculation process was carried out by using the PWV automatic solution platform,and the correlation between PWV and rainfall was analyzed using four rainfall events in Hong Kong in 2019,it can be concluded that the PWV will reach a higher level before rainfall and the rainfall will generally occur in the accumulation stage and the decline stage of PWV.