Research On Theory And Application Of Remotely Sensing Meteorological Parameter Using GPS

GPS meteorology is a rising subject. With the development of Global Navigation Satellite System and LEO Satellite Project, GPS meteorology is rapidly developing. More research projects are provided for us. Research on the projects will help development and application of GPS meteorology.The development status and research achievements of GPS meteorology are introduced in detail, and the research trends are analysed in this dissertation. The composition, characteristics, development status of GPS, GLONASS and Galileo and the course of GPS modernization are discussed. COSMIC, a joint U.S.-Taiwan project and ACE+ which is a project managed by ESA are introduced, including their composition, research fields, objectives, and missions, and their technical characteristics are analysed. The characteristics of ground-based and space-based GPS meteorological remote sensing technique are discussed by comparing with traditional atmosphere sounding means. The theory and realization method of remotely sensing meteorological parameters using ground-based and space-based GPS data are studied. The models and methods are researched by processing and analysing measurement data. The main research achievements in this dissertation is shown as the following:(1) Hopfield model is modified and the reliability of calculating result of using the model is enhanced.Hopfield, Saastamoinen and Black models are main models of calculating zenith hydrostatic delay (ZHD) in estimating integrated water vapor (IWV) using GPS data. Their accuracy will directly affect the IWV accuracy. Therefore, integrated water vapor is calculated respectively by above 3 models using zenith neutral delay (ZND ) and meteorological data in BJFS, LHAS and WUHN stations which have different geographical positions. The calculating results are compared and analysed. The analysing result shows that the IWV of calculating with Hopfield model appears systemic deviation compared with the IWV of calculating with other models in LHAS station. The IWVs that are calculated using Saastamoinen model and Black model are validated by comparing with radiosonde data. Following study proves that the systematic deviation is not influenced by geographical latitude, longitude and date, it is only dependent on the station elevation. The modified termof Hopfield model is first given by fiting the IWV deviation in 23 stations. The checking computation proves that the modified term of Hopfield model can well cancel the deviation due to station elevation, and improve the reliability of calculating result with Hopfield model.(2) The new method of directly estimating IWV using ZND is first advanced. Author finds that the change trends of ZND and IWV are consistent by comparing ZNDwith IWV in BJFS, LHAS and WUHN stations. The new method of directly estimating IWV using ZND is first advanced in further study and analysis for numerical value. The feasibility of the method is testified by calculation of measurement data. If the calculating IWV using former method is taken as true value, root mean square (rms) error is about ±1.0 kg/m2 in not eliminating gross error. The new method of directly estimating IWV using ZND is very useful for estimating IWV in the stations or period that have not meteorological data.(3) The new method of making dynamic modification for transformation coefficient is proposed. The error caused by using different transformation coefficient model and mean temperature formula is analysed.In calculating IWV using ZWD, the key variable is a transformation coefficient. There are two models of calculating transformation coefficient: one is a function of mean temperature, which is widely used, another is a function of station latitude and date. The transformation coefficient is calcualted using two models in BJFS, LHAS and WUHN stations. The IWV bias due to the difference between transformation coefficients calcualted using two models is analysed. The result shows that the bias is relate to station geographical position and date. Sometimes the bias can reach lkg/m2. The mean bias of IWV in LHAS station approachs lkg/m2. The bias is due to model error. So the selection and optimization of the models of calculating transformation coefficient are very important.For analysing the influence of mean temperature, the linear regress formula which is given by Bevis et al. (1992) and the linear regress formula which is proposed by Li Jianguo et al. (1999) are compared. The result shows that the mean temperature calculated by the two formulas is different, and the difference is relate to temperature and humidity. As a rule, IWV bias because of the difference is smaller than 0.3 kg/m2, but it will become large in winter and humid area. IWV bias can reach 0.76 kg/m2 in -20 °C and IWV of 60 kg/m2.For eleminating the influence of model error, improving accuracy of transformation coefficient, a new method of making dynamic modification for transformation coefficientusing rediosonde data is proposed. The method is not validated because of the lack of rediosonde data.(4) The method of assessing the accuracy of IWV calculated with ZWD model by IWV computed using GPS data is first used in this dissertation.The meteorological data in stations includes temperature, pressure, and humidity. The humidity is not used in estimating IWV using GPS data. It is key data to directly estimate IWV by ZWD model. The meteorological data and elevation are used to calculate IWV by Hopfield and Saastamoinen ZWD model, and the result is compared with corresponding IWV calculated using GPS data. The result shows that the accuracy of IWV calculated by Hopfield and Saastamoinen ZWD model is about 20%~30%. IWV calculated by Hopfield ZWD model and IWV calculated by Saastamoinen ZWD model are different, the difference is affected by elevation.(5) How to determine initialization height of the Abel integral inversion using retrieval temperature profile is discussed. The temperature profile change before and after statistical optimization for GPS/MET and CHAMP is analysed.The choice of initialization height will influence retrieval of temperature profile. By analysing the relationship between the boundary condition of temperature and retrieval temperature profile, it can be known that the retrieval result is most reasonable when the upper boundary condition of temperature is most close to the change trend of retrieval temperature profile.By comparing and analysing retrieval temperature profile without statistical optimization and with statistical optimization, it can be found that temperature profile change before and after statistical optimization is obvious for GPS/MET data, and temperature profile change is not obvious for CHAMP data.(6) The effect of gravity value on retrieval result is analysed.When the pressure is calculated, the two methods can be selected: the gravity is integrated in one method, and the gravity is taken as a constant in another method. The temperatures and pressures which are calculated using the two methods are different, the difference of temperature is smaller than 0.6 K, the difference of pressure is smaller than 2 hPa.(7) Exploiting and developing two kinds of software: IWV processing and analysing system and GPS occultation retrieval system.For processing, comparing and analysing large numbers of data, author develops two kinds of software: IWV processing and analysing system and GPS occultation retrieval system.IWV processing and analysing system takes ZND and meteorological data as original data, and uses Hopfield, Saastamoinen and Black models (including ZHD and ZWD models) and two transformation coefficient models. The software can extract the required raw data from many files in format of ZND data and meteorological data provided by IGS besides the function of batch data processing and graph display. It can give more than 30 kinds of data, such as IWV, ZHD, ZWD, transformation coefficient etc for study and analysis.GPS occultation retrieval system takes Level2 data of CHAMP and GPS/MET as original data, and Abel integral inversion is used in refractivity retrieval. The coordinate transformation, ionospheric correction, earth's flattening correction, upper boundary condition statistical optimization etc are used in retrieval process. The software can fulfill retrieval from extra phase delay to bending angle, refractivity, pressure and temperature.