Research On Establishing Accurate Atmosphere Tomography Model Using Reference Station Network Data

Atmospheric water vapor is an important part of the atmosphere, it plays an important role for regional climate, marine, hydrology, geodesy, resource remote sensing, aerospace and other fields. At present conventional atmospheric detection means have many restrictions, but ground-based GPS water vapor detection technology becomes a powerful supplement of conventional atmospheric detection means for its characteristics of low cost, high accuracy, all-weather observation and high time resolution.Sichuan basin is located in southwest inland areas of China, bordering Qinghai-tibet plateau. The weather is changeable, and water vapor is rich. It is a rainfall center in the Chinese mainland away from ocean, and it is also the areas of multiple local rainstorm. Obtaining vapor of high precision, high time efficiency, high spatial resolution is very important for the climate research and meteorological forecast of Sichuan region.This thesis uses the data of Sichuan reference station network to make a series of studies for ground-based GPS water vapor detection technology. The main contents of the the thesis are summarized as follows:1. This thesis determines the strategy of using GAMIT software to calculate absolute zenith total delay of Sichuan reference station network through many experiments.2. Through the analysis the internal and external conformance of three zenith hydrostatics delay models, Saastamoinen model is determined to use for water vapor inversion of Sichuan region.3. This thesis uses experience regression method to establish the weighted average temperature formula of Sichuan -SC formula, and rms is 1.95k. This thesis determines SC formula is suitable for Sichuan region, through comparing precipitable water vapor of SC formula with many other methods.4. Using the proposed calculating scheme of zenith total delay, Saastamoinen model, SC formula and the measured meteorological data, GPS precipitable water vapor of Sichuan reference station network is calculated, and it's average deviation from radiosonde precipitable water vapor is 1.10mm, meeting the accuracy requirement of water vapor inversion.5. This thesis improves the deviation coefficient formula, and illustrates the corresponding relationship between GPS precipitable water vapor and actual precipitation by using numerical computation.6. The pressure and temperature of station is estimated using Global Pressure and Temperature Model, and GPT model is suitable for the network with no actual meteorological data and no high requirement of water vapor inversion, through comparing the precipitable water vapor of GPT with measured meteorological data and radiosonde.7. This thesis adopts double difference method to calculate the slant water vapor, and specifies the method. This thesis selects JYAN, LESH, ROXI and YBIN as reference stations and divides the areas into 6×4×10 tomography grid. Observation equation is established using slant water vapor of 15minutes data of three sessions on February 18,2008, and using Gaussian weighted function as level constraints, using the radiosonde data of the average of five days as vertical prior condition, using LSQR method to solve the equation. Tomographic results may reflect the process of water vapor increasing when rainfall happenning, and water vapor reducing after rainfall, and tomographic contour line is in good conformance with radiosonde contour line, the average deviation is 0.1686mm/km and the rms is 0.1844mm/km.