Study Of Regional Atmospheric Precipitable Water Vapor Based On GPT2w Model

Water vapor,as an important gas in the earth's atmosphere,is not only an important material basis for the clouds and ground rain in the sky,but also one of the key monitoring objects of the earth's climate change and flood disaster.Therefore,it is very important to detect the moisture content and change of the atmosphere.However,because of the complex and changeable time and variation of water vapor sequence and its unstable changes,it is one of the research subjects that many researchers have been studying for the accurate detection of water vapor and rainfall.China has vast territory,large terrain,high terrain,high East and low terrain,various geomorphology,rich and diverse climate types: tropical monsoon climate,subtropical monsoon climate,temperate monsoon climate,temperate continental climate and plateau mountain climate.Therefore,it is of great scientific and practical significance to retrieve precipitable water over China.Based on a detailed analysis of the GNSS inversion of Atmospheric Precipitable Water Vapor theory and methods,based on the high precision and all-weather advantages of the ground GNSS,this paper mainly aims at the problem of the lack of meteorological observation data or meteorological data in some of the current GNSS stations and the current partial tropospheric model of UNB3 m,GPT2 and GPT2 w,which can be calculated according to the location and date of the station.Based on the analysis of the errors of these meteorological data,we put forward the advantage of empirical meteorological data,and put forward the use of these meteorological data to retrieve PWV.Therefore,the main research contents and conclusions of this paper include:1.The current commonly used troposphere model GPT series model and the UNB3 m model in the UNB series model are described in detail,and then,on this basis,the error of the two meteorological parameters,the station pressure and the station temperature,which are calculated by the GPT2 model and the GPT2 w model and the UNB3 m model in the GPT series model,are calculated by the GPT2 model and the GPT2 w model and UNB3 m model.It is found that the accuracy of GPT2 model and GPT2 w model is higher.2.Using the meteorological data provided by 89 radiosonde distributed in the Chinese region as a reference,the error of the classical atmospheric weighted mean temperature Bevis formula in China is obtained and the distribution characteristics of the deviation,standard deviation and root mean square difference of the formula are analyzed in detail in time and space,and an improved model of Bevis is proposed,and the model is verified.The accuracy is greatly improved in different regions,and the weighted average temperature of water vapor is found to be influenced by altitude and influence,but it is also influenced by the latitude and longitude of the station.3.After calculating the meteorological parameters in the resolution precision model of the GPT2 w model and calculating the value of ZWD and K,the comparison between the PWV and the measured meteorological /PWV proves that GPT meteorological data can be used for the meteorological network with no measured meteorological data and the high precision of the inversion accuracy.