The Study Of Water Vapor Inversion Using Ground-based GNSS And Its Applications In Severe Weather Conditions

Though water vapor accounts for a little percentage in the atmosphere, it plays a key role in a series of weather phenomena. Water vapor is the only substance in nature that can have three phase change, which explains the essential reason for various weather phenomena. The phase change of water vapor absorbs or release huge heat, which will exert greatly influence on the vertical stability of the atmosphere, the form and evolution of storms. Dramatic change of water vapor often causes destructive weathers, for instance, typhoon, storm, hailstone, drought and flood, etc., that may bring heavy losses to the security of life and properties. So it is always much concerned to monitor the water vapor change and a good understanding of the temporal-spatial change of water vapor will be helpful to improve the weather forecasts and disastrous weather warning. Additionally, water vapor is the most abundant greenhouse gas in the earth atmosphere and contributes much to the global warming. So, having a good knowledge of water vapor content and its change is very meaningful to short-term weather forecasts, disastrous weather warning and climate change, which is also what state governments, academic institutes and social groups keep working on.In 1990s, the propose of GPS meteorology draw the research interests of many scientists. The GPS technique has the advantages of continuously operating, low cost, high accuracy and temporal resolution relative to other traditional techniques in water vapor detecting. Since April,2013, the IGS started to broadcast the real-time GPS orbit and clock products via the Internet, proving a chance for real-time water vapor monitoring. Motivated by this background, this thesis analyzes the error sources of GNSS water vapor detecting, propose the solution and model for accuracy improvement, assesses the impacts of real-time orbit and clock products on the estimation of positons and tropospheric delays, proposes advanced water vapor tomography algorithms to conduct experiments in heavy rain weather and rainless weather and obtains some interesting conclusions. The main work and contributions of this thesis are as follows:1. The accuracy of GNSS water vapor relates to the accuracy of tropospheric wet delay and weighted mean temperature. By utilizing the error propagation law, this thesis gives the mathematical relationship between the precipitable water vapor and zenith wet delay, weighted mean temperature, determining that the zenith wet delay error and weighted mean temperature error nearly have the same contribution to the precipitable water vapor error. Based on this conclusion, we know that the key of improving GNSS water vapor accuracy lies in the accuracy of wet delay estimation and weighted mean temperature calculation.2. The thesis also analyzes the impact of the error of a prior zenith tropospheric delay on the estimation of final tropospheric delay. As the wet mapping functions are not equal to the dry mapping functions, the error of prior zenith hydrostatic delay will not transfer into zenith wet delay equally, so the final estimated total zenith delay will be affected by the prior values. Simulated data experiment shows that 10 cm zenith hydrostatic delay error will cause 2.8 mm zenith total delay error. When accurate tropospheric delay model is used, the final error can be reduced to-1 mm.3. In order to improve the accuracy of GNSS water vapor from the respect of weighted mean temperature, this thesis proposed different weighted mean temperature models for use under different conditions. Based on the spatial-temporal characters of weighted mean temperature, this thesis establishes the empirical model GTm-III that considers the annual, semi-annual and diurnal variations and geophysical difference. This model has RMSE of 3.2 K and 4.2 K when validated by GGOS Atmosphere data and Radiosonde data, respectively, achieving a leading accuracy among the same type of models and an equivalent accuracy with the Bevis model using in-situ temperature. In addition, this thesis establishes one (surface temperature)/multi (surface temperature and vapor pressure)-parameters based models. The one-parameter model has RMSE of 2.58 K and 3.82 K when validated by GGOS Atmosphere and Radiosonde data, respectively. The RMSE becomes 2.48 K and 3.47 K when it comes to the multi-parameters models. The accuracies of these models are superior to other same type models or empirical models.4. The impacts of IGS real-time orbit and clock products on the estimation of coordinates and tropospheric parameters are also assessed in global and regional scales. The results show that the impacts of real-time products on coordinates and tropospheric delays are within 1-2 cm, indicating the accuracy and reliability of these products and laying the foundation for real-time water vapor monitoring.5. The SVD+SIRT algorithms are developed to conduct tropospheric tomography experiments in real-time and post-processing modes under raining and rainless weather conditions, respectively, and wet refractivity with high accuracy is obtained. It can be concluded by analyzing the results that ①the accuracy of wet refractivity obtained in real-time mode is a little poorer than that in post-processing mode, but the difference is within 1 mm/km, further indicating the feasibility of real-time water vapor monitoring; ②the accuracy of GNSS tomography will not decrease in heavy rain conditions, but will be affected by the vertical distribution of water vapor. These experiments further reveal the great application potential of GNSS tomography in monitoring disastrous weathers like small-range short-term heavy rain process.