ALBERT

Description: Modeling tropospheric delay is critical to achieving high-precision data analysis of space geodetic techniques, such as Global Navigation Satellite System (GNSS), Very Long Baseline Interferometry (VLBI), and satellite altimetry. Numerical Weather Models (NWMs) are an important source to provide both empirical tropospheric delay models and real-time corrections, and are widely used as a priori values, external constraints, and precise correction values. The uncertainty of NWM-derived empirical models is around 3 to 4 cm, and that of reprocessed and real-time products is around 1 to 2 cm. However, this uncertainty is mostly derived using long-term data on a global scale, and the geography- and season-related uncertainty variations have not been revealed thoroughly. In this study, we investigate the uncertainty of NWM-derived tropospheric delay products using a huge amount of globally distributed GNSS stations (more than 10,000) over ten years, and focus on the geographical and seasonal distributions. We demonstrate that the uncertainty of NWM-derived tropospheric delays (1) have clear seasonal variations, that is, larger in the Summer season when the water vapor is more abundant and with rapid fluctuations, and (2) shows a strong dependence on the geography, which is also attributed to the water vapor distribution and variation differences. Our work could provide a reference for future research of exploiting NWM-derived tropospheric delays, especially in real-time GNSS positioning and navigation.