An Inversion Method of Water Vapour Profile by Using GNSS Wet Delay to Get PWV

The traditional method to get Precipitable Water Vapor (PWV) from the Global Navigation Satellite System (GNSS) is by using Zenith Wet Delay(ZWD) by means of the approximate coefficient which can cause larger error when the significant fluctuations of water vapour occur. This paper proposed a new inversion method of water vapour profile by using GNSS wet delay to get high-accuracy PWV. By using ten years'(2014∼2023) historical radiosonde meteorological data at Haikou radiosonde stations, water vapor density profiles are retrieved with the intelligent optimization algorithm named Relevance Vector Machine (RVM) based on high correlation between ZWD and water vapor density profile. The highaccuracy PWV is then integrated by water vapor density profiles. The PWV computed from water vapor density profiles are compared to the PWV from the traditional method, the results show that the RMS of PWV computed from water vapor density profiles agree to within 1 mm which is much precisely than the traditional method, which enhanced the precision and versatility of PWV.