Experimental Research On Retrieval Model For Precipitable Water Vapor Based On Multi-source Data

Precipitable water vapo r is an important greenhouse gas and affects atmosphere’s stability directly. Space-time distribution of precipitable water vapor is important to disastrous weather forecast, climate research and other related to water vapor application. At now, there are many methods to get precipitable water vapor, but all of these methods have advantages and limitations, so we must make use of all methods integrated to achieve complementary advantages on the basis of full research on every method. So precipitable water vapor’s precision, time resolution and space resolution will be improved hugely, it is convenient for research and application on space-time distribution of precipitable water vapor.The retrieval methods of precipitable water vapor and modeling based on multisource data is researched in this thesis. In this paper, on the basis of studying four kinds of data retrieval methods and analyzing five kinds of methods’ characteristic and precision, also including estimation by ground humidity parameter, we modeled precipitable water vapor retrieval model based multi-source data.Firstly, studying precipitable water vapor retrieval theory and methods of radiosonde observation, ground-based GPS,near-infrared of MODIS and near- infrared hyperspectral of HJ-1A.The technique and steps of installing and configuring GAMIT which is named high precision GPS data calculation software on Ubuntu operating system are summarized in the chapter of retrieval theory of ground-based GPS observation.In the part of retrieval theory of MODIS,we also achieve cloud detection.Modeling, analyzing and choosing bands and relation simulation between atmospheric transmittance of water vapor absorption bands and precipitable water vapour are performed in the chapter of retrieval theory of near- infrared hyperspectral of HJ-1A.Precision validation and analysis have been done about retrieval precision of radiosonde observation, ground-based GPS, near- infrared of MODIS and near- infrared hyperspectral of HJ-1A at the study area according to retrieval theory and analysis. We used ultra-rapid ephemeris(igu) and final ephemeris(igs) to validate precision respectively in the chapter of validation of ground-based GPS.Then, we corrected near- infrared of MODIS retrieval results by ground-based GPS retrieval results and build liner regression calibration model and near real time calibration model to improve the retrieval precision of near- infrared of MODIS. Near-infrared of MODIS retrieval results are used to compare with near-infrared hyperspectral of HJ-1A retrieval results in the section of retrieval precision validation of near- infrared hyperspectral of HJ-1A.According to retrieval results of multi-source data, we summarized the characteristic, advantages and shortcomings of every data source. Then we build retrieval model based on multi-source data. Retrieval model based on multi-source data could supply high precision, high time resolution and high space resolution precipitable water vapour data and lay foundation for precipitable water vapor’s time and spatial distribution research.