The New Improved ZHD and Weighted Mean Temperature Models Based on GNSS and Radiosonde Data Using GPT3 and Fourier Function

Compared to the zenith hydrostatic delay (ZHD) obtained from Saastamonien model based on in-situ measured meteorological (IMM) data and radiosonde-derived weighted mean temperature (Tm), ZHD Tm deviations of GPT3 have shown obvious periodic trends. This article analyzed seasonal variations GPT3-ZHD GPT3-Tm during 2016–2020 period in Yangtze River Delta region, new improved models were established by multi-order Fourier function. The precision improved-ZHD was verified using IMM-ZHD products 7 GNSS stations period. Furthermore, precisions precipitable water vapor (PWV) PWV 2016–2019 with GNSS-PWV products, Bias RMS are −0.5 mm 2.1 mm, while those GPT3-PWV 2.7 11.1 mm. Tm, −0.8 K 3.2 K. 2019 2020 −0.1 0.5 respectively, decreasing 0.4 1.6 compared GPT3-ZHD, improved-Tm −0.6 K, 0.2 GPT3-Tm. calculated GNSS-ZTD, improved-ZHD, 0.6 GNSS-PWV, 2.2 10.5 GPT3-PWV. It indicates that can be used obtain high-precision PWV. applied effectively retrieval real-time region.